Skip to main content

Full text of "A Rudimentary Treatise on Clock and Watch Making: With a Chapter on Church Clocks; and an ..."

See other formats


This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing tliis resource, we liave taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for in forming people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liabili^ can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http: //books .google .com/I 

le JLitiratg 








\^ nnmerotis IStatomgs* 









T-e-C 5 ) V -b-'o 


^mvj^^L OF ENeiJEjaim' 

JUN 20 1917 

t • i--^* ■ 1 1 

A , ' 










. 173 


Foiar-wheeled train (drawing) 

. 175 


Three-wheeled trains .... 

. 178 


Doable frame (drawing) 

. 180 


Jack wheel for winding . . . . 

. 181 


Long barrels 

. 182 


■ Wire ropes 

. 183 


Double hammer 

. 184 


Calculation for three-wheeled trains 

. 185 


Striking part for three-wheeled train 

. 186 


Shape of crutch 

. 188 


Spring fork 

. 189 


Dial and leading-off work 

. 190 

Friction rollers (drawing) . . . , 

. 192 


Adjusting work 

. 193 


Size of dials 

. 196 


Material of dials .... 

. 199 


Illuminated dials .... 

. 200 



. 202 


Maintaining power for winding 

. 204 


Bolt and shutter .... 

. 205 


New construction of (drawing) 

. 209 


Mr. Airy's going barrel (drawing) 

. 211 


Striking part (drawing) . . . . 

. 214 


New Striking wheel (drawing) 

. 218 


Loss of power in striking . . . . 



Quarters (drawing) 



Chimes of St. Mar/a, Cambridge . 



Bells . . . . 







Time of striking first blow of the hour . 



Bemontoire in train . . ■ . 



Bojal Exchange dock remontoire (drawing) . 



iVench gravity remontwe (drawing) 



Bemontoire with continuous motion 



Spring remontoire, Meanwood clock (drawing) 



Cast iron wheels ..... 



Gou-metal ...... 



Painting and oiling 






Case for clock 



Provision for weights falling 



Proceedings respecting the great Westminster 




Mr. Airy's conditions ..... 



Pendulum accderator .... 



Mr. Airy's report 



On public docks in general 



On specifications and tenders for docks 



Church docks without dials 



Advice to churchwardens .... 





Page 11, line 5, after ^ibr read the day before, 

19, ,; 5, iot first read innermost , and for make read mark. 
6, for «, read «s. 

5> »> 


76, „ 3, for - read -r . 

„ — „ 5, for -^ read -^. 

3> 77, „ 5, for lose read ^tfw. 

,, 92, „ 7 from bottom, for groce read groote. 

„ 132, „ 9, after thai insert o». 

„ 169, „ 5 from bottom, for is the only one, read and the 

Cambencell church clock, aUo by Mr. Dent, are 

the only ones. 
" 202, „ 3, after dial insert and one-fifth of the radivs is 

large enough, 

224, „ 4 from end, for required read requisite, 

225, „ 7,ioTj!)eelTesidpeal, 
250, „ 5, for it read they. 



I have several times been asked if there was any book on 
clock making, which, withont being encumbered with profes- 
sional details, would enable a person having a moderate know- 
ledge of mechanical principles to understand the construction 
of clocks, and the things which chiefly require attention in 
their making and management. I believe there is no such 
book. Perhaps the articles in some of the Encyclopsedias 
might partly answer the purpose; but they are neither 
accessible to the great majority of readers, nor do they 
contain some information which may be expected in a 
rudimentary treatise, though they do contain iuteresting 
accounts of some curious inventions, which would be out 
of place in a merely practical book such as this professes 
to be. 


On the subject of chxireh, or turret, or public clocks 
especially, there is almost a complete absence of literature. 
And yet this is just the branch of the art on which, I am 
sure, from what I have seen of such clocks, some general 
information is most needed, not only by the public who 
have to pay for them, but, I presume to say also, by many 
of the clockmakers who undertake to construct them. 
Conmion house clocks, like watches, have become merely an 
article of commerce ^^an&liheotdy^oBiem is how to make 
them to sell. Astronomical clocks or regulators, on the other 
6and, are only made by the^hesLmakers, and the attention of 
scientific men, both professional clockmakers and others, has 
been chiefly directed to their improvement; and when it is 
stated that a gbod clock of this Mnd can be depended on 
ib one or two seconds in a week, it wiQ be admitted thai 
that attention has not been fruitless. But with church and 
tuwet clocks the case is quite otherwise. They have been 
regarded merely as common clocks on a large arid coarse 
scale, and it has therefore been almost taken for granted 
that one clockmaker cotdd make them as well as another. 
Whereas in fiict their largeness and heaviness is just the 
thing that makes them difficult to make well : a chclc, at 
least the |ime-keeping part of it, Being the only machine 
whone sole business is to overdone its own friction ; and there 
is greater room for difference in their construction and 
value than in any other kind of clocks. 

• Hk' ^antitiwigifc^ which ^wg ;aEe '^ofiMTBied ia esti- 
-satui^ iteeObolB of/ aifepemeiili iaiid^ tifieiif (hidaob 

•of finlMHi iltfdf is tsdr entkcify 'eiap^;fes!> ^Ml tedly good 
idackt «ai ntrar be mBJi^- ^irtlbdttl -s^td^ ^ combination 
t<if MmAmiaJfcioal '«imI |fi!M^tk»iI lEk)wledgl^ its of course 
isw doAmskeH cam be H^icpk^tdi ' fO ]podses8. I hope 
Ad tUs 4f6a(lise ntty tccasiO!!]^ seMcoaUe to 

.thMi-ia ilttis res^eot^ aittt 'wfth the ^rieftr of wsikmg It 
miro *a r bfcre not schijieS, ^iR^ieHBTter any prac- 
tical results -ctep^aidtod on it^ t6 pttt down md;t)ienudicd 
results in their most general form^ although in several 
faistances the investigations, by which very simple results 
are obtained, are too complicated for any attempt even to 
Indicate the nature of them in a rudimentary treatise such 
as this : I refer especially to the method of calculating the 
disturbances of pendulums, given by the present Astaponomer 
Boyal-in a paper in the 'Cambridge Philosophical Trausr 
actions/ vol. IIL, of which the substance may be seen in 
' Pratt's Mechanics/ I may add that it is no detraction 
from thje value of those calculations and the formulse obtained 
by them, that they aSoati the means of arriving at ^ome 
practical conclusions respecting the capabilities and the xaU* 
tive merits of the several closes of escapemepats different from 
those which were stated by Mr. Airy hnoself, and which 
1 have accordingly not thought it necessary to mention. 


I have also to express my obligations to two eminent 
dockmakers, Mr* Dent and Mr. YuUiamj^ for much prac- 
tical information which they have readily communicated to 
me for the purpose of rendering this little book more valu* 
able, both to professional and general readers. I must add 
that it is due to Mr. Denf s enterprise and determination to 
find out the best way of doing everything connected witb 
his art^ that I am able to state the actual results of 
some experiments (of course expensive in the first instance) 
which he ventured to try at my suggestion la the two 
church clocks I have several times referred to. 

From the interest which the subject of the Great Clock 
for the new Palace at Westminster excited in both Houses 
of ParUament three years ago, as weU as from the many 
questions and observations I have since heard about it, 
I believe that the short account which I have given of the 
proceedings respecting it up to the present time will be 
acceptable to many readers. Whether that clock is ever 
really to be made is a question which requires some 
other science than that of horology to predict. The 
government pledged itself to the undertaking by 
the commission which it gave to Mr. Barry (at his own 
recommendation) six years a^o, as well as by the repeated 
assurances of the past and pr^ent Eirst Lords of the Woods 
and Forests, that ' all the resources of science were to be 


employed upon the Great dock/ But from the time 
when. Lord Morpeth gave that answer until now^ nothing 
wbatever has been done in the matter. I understand that 
there are alwajs torrret clocks exhibited at the French 
EqK)sitions; and as this is intended^ we are told^ to be the 
best and largest clock in the world, nothing could be a more 
appropriate subject for exhibition, and for reward if made 
as it ought to be, at the first exposition of manufactures for 
the world, than such a clock ; to say nothing of the real 
want of it as an accurate public 'regulator,' accessible to 
everybody in London, either by sound or sight. 

Perhaps I o«glit to apologise for occasionaUy expressing 
opinions of my own in a manner which may appear unusual 
in treatises of this kind. But several of the subjects I have 
had to treat of are so much more matters of opinion as to 
the balance of different advantages, than of demonstration 
or of fact, that the reader would have had much greater 
leason to complain, if I had merely stated just as much as 
is ascertained, without giving him any assistance on more 
doubtful matters ; and of course I cannot pretend tq think 
that, if I am qualified to write a treatise on clock-making at 
aU, my opinion on such matters is not worth stating. 

I hope especially that the suggestions I have offered for 
the assistance of churchwardens and others who want public 



olocks^ may (naoaaiottallf kmsskat geGAdoA,iwsb^ ei r^^M 
oi)^ to bQ ^rtylaiuad for Ab Inll■ie7^iiiqr have U sp^ifl. If' 
the book, elfeete this, or lany ethnr ctf the objeels I ha^iSf 
idS^rel to, I dtalt besdlbfied iliat iotjMtn^ (Moml Beif s 
isecommondlitiaii io thq pfeiUiBlief^ "bf tbss addiidon tOf the 

bftdoes>Lsu8riD0itbeeiLtaiuplfioid«' - . / 


I • . • . . '. 

E. B. D. 

' .-. ; ' 

" ( . "1 

42; Queen Anne Street^ London : 
1, Jan., X85d. 

, » 

■ •.• ' 1. ' '. -•' 


,;; i iNTK0ti^>U€TrON J 


, • • • - ' • 

■' i. B«Pb!KE ire isxamin^ ^he constraction of instruments 
to ^aisttre' tinie, it' "will ^t W veH io niiderstanS cleaffy 
whdtlt fs <Rat w6 Trant to lireasmre. • < 

'' ^.169. 6l a ceirtaW lei^fli is called a yard, und is ow 
stafaiiiM (#ltegtH;''lyttt it ililglri; fast as well have teen 
airjr 'otter 'ferigft.' "Thieife 'is'no stch thing as a natural 
yard, which we aiti Iqtrite mire'wMieVer we meet with it is 
the thing we call a yard, and with which we intend our 
sifenllard'- to agife^i "'It Is plirely afbittary, likje a pound or 
ai>galk«i.'.^ Aia^t* is^rtOt'eVeij^Wify that if afl 

our yard-wands and other measures were burnt, they could 
opfj iia re^pr^d bj^th^^j&coi^ ti^ t^^y bose a eertapi. 
proportion, (nofy ^etpiPesriUe'^ h Ibng ^(mi^t k^ igoreM, • ta 
tK^ lenirfh' of "a periduluni *diat witt vibrate^ certain num- 
\^ of. times, ita ai(^j|^aiUi:laiijta4^.0nd:At^a^ 
ture, during one revolution of ftief earth/ ^ i * • - 


This revolution of tlie earth upon its axis is the onlj 
natural measure or standard of our time. But it is not 
what is commonly called a Day^ which is generally understood 
to be the time between two successiye noons or midnights* 
But even thia is not the day of twenty-four hours by the 
dock^ consisting of so many minutes and-seconds. Let us 
see then what a day really is» 

2. If you fix a telescope on an axis, so that it can only 
move in the north and south plane, and observe the time of 
any fixed star passing the middle of the telescope, that is 
the meridian of the place, twice ; that time is the exact 
period of the earth's actual revolution, without reference 
to the sun or any other body, and is what we call a sidereal 
dfiLy^ and is always the same ; and this kind of day, with 
its subdivisions of sidereal hours and minutes, is used for 
many astronomical purposes, and occasionally for correcting 
the artificial time shown by ordinary clocks* 

3. But if the telescope (which when so mounted i» 
called a transit instrument) is directed to the sun, instead. 

* This is not strictly true, for sidereal time is really measured 
from the transit over the meridian of a certun. imaginary point or 
line, which is the intersection of the equator and the ediptic, and is 
called the first point of Aries T : and this point has a yeiy slow 
preeestum, amounting to 50" in a year ; which, the reader may be 
reminded, is a veiy different thing from 50 seconds of time, being iir 
fact equivalent to '6\ seconds of time. 


of a star^ it will be fouad that the sun takes 3m. 56* 5554s. 
of sidereal time longer^ on the average^ to pass a second 
time than a star does. The reason is that the earth has 
been moving in its orbit in the same direction sus it turns 
upon its axis^ and therefore^ before the same place on the 
earth can i^ain look full at the sun, or have the sun on its 
meridian, the earth must tiim rather more than once round. 
The time of this passage of the sun is called a iolar day, 
and it is evident that in the year there must be one less 
sokr day than there are sidereal. 

4. I said the solar day was longer than the sidereal by 
3m. 56^8. ofi the average. Eor it will be found on accurate 
observation that this difference is variable. About the 10th 
of February, the 14th of May, the 25th of July, and the 
2nd of November, the difference between a solar day and a 
sidereal one will be very nearly constant for several days ; 
but in the middle of April and June it wiU vary from day 
to day about 11 seconds ; and at the beginning of Septem- 
ber and at Christmas, as much as 20 and 30 seconds ; and 
in February and November these variations have accumu- 
lated to so much that the sun appears to be on the above- 
mentioned day in February 14 minutefs too slow, and in 
November 16 minutes too fast; and about one-third as 
much respectively in July and May. This variation in tke 
length of the solar day is caused by the elliptical form of 
the earth's orbit* However in a treatise on clocks we are 


only concerned with the fact, not the explanation of it, for 
which astronomical books must be consulted. 

5. Since the sidereal day does not suit our ideas of day 
and night corresponding to light and darkness, and the 
solar day is of variable length, a third kind of artificial and 
uniform day has become necessary, now that all the time of 
the world is measured by clocks instead of sun-dials. The day 
which is so used is that which is always the 3m. 56 '5554s. 
of sidereal time longer than a sidereal day; and this 
artificial day is called a mean solar day, or more shortly, 
a mean day ; and the time shown by clocks is accordingly 
called mean time,* And the difference between mean time 
jtnd the time shown by a sun-dial or any other solar in- 
strument is called the equation of time, which as we have 
seen is sometimes as much as fourteen and sixteen 
minutes. The equation of time is not quite the same for 
the same day, in any two successive years, on account of the 
leap-years, but it differs so little that a table to the nearest 
second for every day in the year 1850, will be sufficient for 
all ordinary purposes, and I have accordingly given a table, 
over the next leaf, that it may all appear at one view, 
of the time which ought to be shown by the clock when the 

* * The astronomical mean day is not reckoned from midniglit, as 
tlie common or civil day is, but from the following noon, and has no 
' A.M.' Thus 11 A.M. May 1 ' civil reckoning* is, in * astronomical 
reckoning,' 23h. April 30. 

3HT >ioi :lihat 



.:j< y. 

•1 ' // ij o 

f. f 

6 .'j*,: A V / 

sun is on the meridian, which is a very convenient form pf 
;Tli§^.table for.pfjffmioatti^, 4si1;i.»fiire8,»Ii;,aftlciil»tioai^ In 
I oring the tal^ tpteonine the gqinj; of , a c^odt between any 
=^<Miaj£(;bQfote 3all:afl«r a liikrof dE'ebtnAty^yott ninsl tise 
jflfe%rrcijttDn giv^n*'i4 theit^Je; for* ^he'^tter Qf 'the two 

I i- ' ', 

' ^:'-. ^ 

V. X take £h]^ op^oi^iuiity o|riniii^rtiiig^ short} iaiih of the 
ififfcgtenc^ bet^en|©i:e^nwicjlttiine and the local time of a / 
jKw Qonsiderable Trapes*: — ' , ; .. ' " 

I V 

."S 3 




1 ■ ■ 
-^BWninglii^.;;; :.:.(. ^ 7 33' 

Chester [] j.l'... ' 11 d&' 

E9inbiirgK....»^« >..».. . JL2 4S 



t ''f <:■ 

^CimbriclgtJ .•../. 


^, xipieter ^.«»r...»^f..».i*i,^... ,. yL4 liJ^; 


Dover ..'. 
, Korwich 


I Boston ,„.,. 

' Gksgow ..../.; ,wr..'l7 

Hnn ...;.„,..... 1 S 

Lteds * 6 -* 

■ liverpoQl ......^..... 11 S3 

ll£aiichester .».« ». . 9 

Newcastle 1 6 34 

Oxford ^ «. ^... 5 

BftrtamooUi ^. ^... 4 24 

Ywk 4 34 

SI. James's, Piccadilly ... 33 

: ■; . ^ 


Qtimsby ...;.. 
Ijputb '...ti..^. 

Boris. ..«« 

Home ......... 

y^emia .*,....»». 

Petersburg ,.. 
JTanJuort ...«»« 



■ 28 
3 32 
' 5 16 
5 12 

9 21 

49 54 

65 32 

121 1 

115 40 


53 8^ 
24 87 


Showing the !nine wbiA. a dock sjioald indicate when the Son 








H. M. 8. 

H. M. 8. 

H. X. 8. 

H. X. 8. 

H. X. 8. 

H. X. 8. 


8 51 

18 54 

12 36 

8 59 

11 66 57 

11 67 26 


4 19 

14 1 

12 24 

8 41 

56 50 

57 84 


4 47 

14 8 

12 12 

8 23 

56 43 

67 44 


6 14 

14 14 

11 58 

8 5 

56 36 

67 64 


5 41 

14 19 

11 45 

2 47 

56 80 

68 4 


6 8 

14 23 

11 81 

2 29 

56 25 

68 14 


6 34 

14 26 

11 16 

2 12 

56 21 

68 26 



14 29 

11 2 

1 55 

56 17 

68 86 


7 25 

14 81 

10 46 

1 38 

56 13 

68 48 


7 50 

14 32 

10 81 

1 22 

56 10 



8 18 

14 82 

10 15 

1 5 

56 8 

59 12 


8 87 

14 32 

9 59 


56 7 

59 24 



14 31 

9 42 


56 5 

59 86 


^ 9 22 

14 29 

9 26 


56 5 

69 49 


9 44 

14 26 

9 9 


56 5 



, 10 5 

14 23 

8 52 

11 59 48 

56 5 



10 25 

14 18 

8 34 

59 34 

56 7 



10 44 

14 13 

8 16 

59 20 

56 8 



11 8 

14 8 

7 59 

59 6 

56 10 



11 21 

14 2 

7 41 

58 53 

56 18 

1 6 


11 38 

18 55 

7 22 

58 40 

56 16 

1 19 


11 54 

13 47 

7 4 


56 20 

1 82 


12 10 

13 39 

6 46 

58 16 

56 24 

1 44 


12 25 

13 30 

6 27 

58 4 

56 29 

1 57 


12 39 

13 20 

6 9 

57 53 

66 35 

2 10 


12 52 

13 10 

5 50 

57 '42 

56 40 

2 22 


13 4 

12 59 

5 31 

57 32 

66 47 

2 85 



.. 12 48 

5 13 

57 23 

66 54 

2 47 


13 26 

4 54 

57 14 

57 1 

2 59 


13' 36 

4 36 

57 6 

57 9 

3 11 




4 17 


57 17 


ii OD the metidum, fot ereiy Day is the Tew 1S60. (Sw j G.) 







H. M. ■• 

H. M. «. 

H. K. ■. 

H. «. «. 

M. K. ■. 

H. K. ». 


8 28 


11 SO 84 

11 49 43 

11 43 44 

n 49 18 


8 83 

8 67 

GS 85 

49 24 

43 43 

49 88 


S 40 

S 83 

SO 18 

49 a 

48 4S 


, * 

8 B7 

6 48 

88 87 

48 48 

48 44 

50 24 


4 8 

6 42 

88 87 

48 80 

48 46 

50 49 


4 IS 

8 SG 

ES 18 

48 12 

48 48 



4 28 

B 80 

57 58 

47 SB 

43 51 

51 40 


4 88 

8 23 

67 88 

47 88 

48 65 

82 8 


4 47 

B 16 

S7 17 

«7 22 


82 88 


4 58 

8 7 

66 57 

47 6 

44 5 



S 4 

4 SS 

S6 86 

48 61 

44 12 

f3 28 


8 12 

4 48 

56 15 

48 88 

44 19 

83 88 


5 SO 

4 88 

68 64 

46 21 

44 27 

84 24 


6 27 

4 28 

56 88 

46 7 

44 38 

84 63 


5 84 

4 18 

B6 12 

45 54 

44 48 

6S 22 


G 40 

4 5 

54 51 

46 41 

44 87 

H 81 


i 45 

8 E2 

84 80 

48 29 

45 8 

68 20 


5 EO 

8 40 


48 17 

45 21 

88 SO 


B 56 

8 26 

B3 48 

48 6 

48 84 

67 19 


B 6S 


53 28 

44 65 

45 48 

57 4Q 


6 2 

2 58 

88 8 

44 45 

46 2 

68 19 



2 43 

B2 44 

44 88 

48 18 

GS 48 


fl 7 

2 28 

E2 28 

44 27 

46 S4 

69 18 


8 a 

2 12 

82 8 

44 20 

46 52 

59 48 


8 10 

1 56 

El 42 

44 IS 

47 10 



8 10 

1 40 

61 22 

44 6 

47 28 



6 10 

1 23 

61 2 

44 1 

47 48 

1 18 


8 9 

1 8 

Btt 42 

43 86 

48 8 

1 47 


e 8 


50 22 

43 52 

48 29 

2 17 


6 8 


50 2 

43 48 

48 51 

2 46 





6. There is however an easy method of testing the going 
<]|f ^ QJopk fcQ3?q^ sid^Te%l^jDil?s»ryi^09|i,,isd^ r^pJ5en(» -j 

tjfrfee eqnatioB ^ tiite. -dbij-person of oi Jinai ' j u n d er stand -- - 
it^g: wd d^tecitj si^y. ;.cqiisti9ia;fci $ iranpit .itu^npaooiLt^f 
sufickttitly cxaJcil for^^thfs purjfosfe^ o^;.^^! a verficil sljb' 
or mai^ in % ^^ed iwiA(j!9w^ aod ^the edge. of a -^hjpmejf^ , 
pino^ded id k iipriglit> aiid the 4sJit- lio plicecf as tp te in thfe " 
mciidi&n of 4hi^ chi^fie;^ - edge. But :it will nttilrally hp" 
aeied how i? the slit to be rb-de; in the jneridian.! Fojr ' 
ofajmasy astconemy thd fpiUowing methoid wiU be jsufficientlr ' 
apcnrate. Kx iipon'a vtrtioai'edge 4f'a cTiinmqr dn the ' 
sotlth ef sotiae *iHndpw^. ^^Detei*Biiie fiifet^of all,! by seeing ^ 
wh^&the shadow of the.3Tin {^Ila>^t ^(^aioon^^or^herei- 
abokit*) which pane ^rfll fce theproper one fbr the- |)ittpos^, ' ' 
and* get the dock set as tiearly as you can to thelright tim^, ' 
eilSierby tradiden from idme persbii who jpdssessiBS^thjs traje 
tii»ie, or from an accvqrata sun. di^ Bod in anj^Jipianak; 
which gives «uck thiiigs ^ftie tinie trf soufMififf, or pasang of ' 
tjie meridian, of any ^bwn star or planet which happens to he 
visible at a coiivenient time fallowing for your longitude, 
if any> from Crreeprcicii). Whai the ^ dock is near the 
tjnie at which tiie star-ought to be on t^e jneridian, stanjd' 
by the window with a sheet ot tin made exactly rectangulair 
Ki^ capable of sliding al^g iihe pane, re(rtiQg on the bar of. 
the window (s«5)popiiig it to *e horixontal). When the 
<3lock is exactly at the time of southing of the star, yoa 
must place the tin so thai you just see the planet past the 


edge of the chimney and of the tin ; and there the tin mnst 
be fixed, or the glass marked and painted up to the 

If the edge of the chimney was its western edge, the 
planet will emerge from behind the chimney ; and then, if 
the glass is painted on the western side of the vertical mark, 
by putting your eye nearly in a line with the mark and the 
chimney edge, you can enclose as fine a line of light as you 
please between the edge of the paint on the west side of the 
meridian and the body of the chinmey on east side, or vice 
versd; and that fine line of light is the line which stars have 
to pass when they pass the meridian. Before the mark 
is finally fixed, the operation ought to be performed several 
times, as there is sure to be some loss of time at the first 
trial in getting the tin to the proper place. You must 
observe the time of transit by listening to the number of 
seconds from the time when you looked at the clock. 

Now when you have got this meridian line, even if it is 
not perfectly exact it will do very well to try the rate of the 
clock by, though not to tell the absolute sidereal time, 
unless it is either quite exact or you can learn how much it 
is wrong. To try the rate of the clock, observe the time of 
any well known star passing your meridian; and go and 
observe it again at the end of one or more sidereal days. 
Your clock ought to be behind its former time, at the 


rate of 3m. 55*9093s. of mean time for every sidereal day 
between the two observations. 

7. Still we want to know how to ascertain the actual 
mean time from a sidereal observation. Now or 24 
o'clock by the sidereal clock of any place is^ as was stated 
in the note to p. 8, the time of the imaginary point r 
passing the meridian of that place. And from this point 
the right (ucefmon^ or distance along the equator (corres- 
ponding to terrestrial longitude) of all celestial bodies is 
measured; and it is generally expressed in almanacs in 
sidereal hours and minutes. Therefore in fact the B. A. 
of any star^ given in the almanac^ is the sidereal time at 
which it passes our meridian, since r was on the meridian 
when the sidereal time was 0. But the mean time was 
when the mean sun was on the meridian; therefore the 
E. A. of the mean sun substracted from the E. A. of the 
star (with twenty-four hours added if necessary) is the 
number of sidereal hours, &c. that have elapsed from mean 
noon to the time of transit of the star ; and this number 
of sidereal hours, &c. may be converted iuto mean ones by 
multiplying them by '99727, the ratio of a sidereal to a 
mean day; or if you substract 9*83 seconds from every 
sidereal hour and 1 second from every 6*1 minutes, it will 
give you the time which ought to be shown by the mean 
clock at the transit of the star. In some almanacs* the E. A. 

* One that rejoices in the singular name of Zadkiel's Ephemeris 


of the mean sun at noon is given for every day under the 
head of sidereal time merely. If the B. A. of the real sun 
only is given, that of the mean sun is found by merely sub- 
tracting the equation of time when the clock is behind the 
son, and adding it when the dock is before the sun. 

8. The more obvious way however of finding the error 
of a clock is to observe the time when the Sun passes the 
meridian ; and if the time then shown by the clock does not 
correspond with the equation of time, the clock is so much 
vrong. The sun is too large to be observed directly like a 
star at the moment of passing ; but it may be observed by 
the same method as I have described for a star^ if the times 
of its western edge first appearing and the other edge dis- 
appearing be observed (with the help of smoked or coloured 
glass, or a fog), and the bisection of them taken for the 
time of noon. Various kinds of meridian sun dials may be 
used, and no others can be very accurate on account of the 
refraction, nor can they be so easily made. If a narrow 
vertical slit is made in any thin plate &cing pretty nearly 
south, the bright line that comes through it from the sun 
will always fall in the same place at noon on any surface 
which is level in the east and west direction, and it can be 

gives it, for the special purpose of saving its constilters the trouble of 
adding or subtracting the equation of time. It may of course be U3ed 
for horological purposes without the necessity of believing in the 
more abstruse science for the benefit of which it is published. 

If) imaQPVotioQE.; 

deasribo bor-ftiAeiifliaiiriiibik.oE iUi Und nn^ b&niide' 
wtAeat toy. eahknaossAsn^anee;. ' < . ,- - '■ i '.'•' ■ 

Make ■ Hmall Imle da a thm' fi»i» of'pMated'tai.'Or I 
brasSj set facing the south in a window-frame or other I 
fltmmBfl&t piaee.. Flare a P^bss «f -ikte or b smooth board 
4|Kt« knfil i^dnsi the< jdbto, Ukd Abeot 9 inches bdow Mtb 
bole. I 9B5qnitel^(^^iecatH»ituea«flrionnk£FitleV«l 
in Hi«[7 ^recticm thiiL,t»'£lid -out io-t^ -flitt iat^i^oe vMtft' 
it the eastand west direction. Sot after the meridian it 
fitrtmd, the bond nay^bfl sloped i^loimi^ the semth, sat 
tikCAhadoiTTiUbs idcaiat^'-aad ehsipei tbra if M is' }evd^. 
Morenrrai, tlubnght^^denrof ithe htde-ii^ b«''cl«u«t< S 
fhe plats is butw is tote puiUcl totiwen^fj asis;>1f 
itinakes an u^ of »b4al 4&? with'tbe^verticallt WiQ-^ 
fnita i the'tole'iii 'tlie 
' plate drop 'a perpeiidi- 
feular A ij upon "t'he 
board^ and'iiMe a mart 
at C. it a little after 
11 o'iJIock on a feie 
day, observe where the 
image of the Sun falls 
on the board, and mark 
the pla^^e of ita ceotre. 
Su and dnw & iftrele on the betfd with tafias &«,. 

Qbsenre'the pbee of ike bii^vpetvii imo ddnei times before^ 
noon^.and yuA. {faemJft,, 83^ and dnnr circles far IkcOft m 
lakamaaiifirw ThoD^if thelaiikcbservsttaiiirBBittllb. 46^ 
go agam a> little before 1^&« Mm., and \99^teb lov tbe pas** 
81^ of tiie 9po4 over r^be ftf«l ^iireH and make the ^aoe^ 
vbere the enatre eioates^ ft^ 0^^ and do tbe «M» with tii» 
other circles. Bidi^t 1^ thi^^ >atei» Si 8x> % s^t, C^ Sjj mxt- 
if you can draw a straight hne through all the bisections 
aad tbe. goint C^ ^t &ie B C is tte in^ffim Jine on 
whibh tine image Bf i^e sita wiffi ^ajis ft!Q at solar noon/ 
«ad the padeiig^ fi ihe )spbi& can be dbsened rtexAf to a 
aetecmd. T!d» lAerdbre n4tli 4he equattE>n4»ble wifl giver 
the meaiw of ^tteetifig a doek on any feie' day. . Befttf^ 
<he Ene is strongly uiarKed it vfM be as w^ll to try it both 
iii flntmmcfr ted at the disfcfaee ot thiree' or fottr months. 
This apparatus is so fcfcsily made, and so nseful irien it is 
made, that eveiy dockmaker who has not the means of 
dbtanring the thrie from an observatory 6ng!it to hare one, 
for it is no nse having clocki? tiiiless you ean contmually 
recnr to wliat maybe called the daium line, from which they 
are to measure, viz., mean noon. Ihere are many large 
towns in which there are literally no means of recovering 
ffic tnte time whek it i lost except by sending somewhere 
else to f^tch ft; £hid th^re arcj very fefw people who are able' 
to take it- aGturatd!y when they ^Q^\t, " -"■''• 

* %, 5Chereis besides IKig simple kind of meridian dial,* 


which anybody can make^ a very beautiM little instrament 
invmited a few years ago by J. M« Bloxam, Esq.^ of 
Lincoln's Inn^ called the dipleidascope, from which^ laien U 
u once accurately fixed, the time of solar noon can be 
observed with even greater accuracy^ on account of the 
increased sharpness of the images which it produces^ as well 
as their passing with considerable velocity. 

It presents Wo bright images of the sun, which 
approach each other, and exactly coincide at noon; and 
thus two observations of noon are obtained, viz., the time 
of first contact and of final separation, the bisection of 
which gives the time of noon, and the time of complete co- 
incidence. It is of no use to describe the construction of 
the instrument, as few persons could make it, even if they 
could legally do so ; Mr. Bloxam has transferred the patent 
to Mr. Dent, who makes them, and from whom also Mr« 
Bloxam's pamphlet on the principles and construction of 
the dipleidoscope may be had, as well z& the annual equa- 
tion table, which is for this year that given a few pages 
back, to the nearest second. 

10. I have no intention of inquiring into the history 
of the various instruments that have been used for horo- 
logical purposes; Sun-dials, at least for noon, could not 
fail to be very early used, from the obvious fact that the 
shadow of any vertical body always Calls in the same place at 


the time when the length of the shadow was least in every 
Jay; and it would be easily observed that the intervals were 
very nearly eqnal. The sun-dial of Ahaz will occur to 
every one. But probably much more accurate astronomical 
observations than any that can be made with a sun-dial 
liad been made long before that time by the Chaldeans^ for 
without accurate measures of time astronomical observations 
can hardly be supposed to exist* 

11, Everybody has heard of the clepsydras^ or water- 
clocks of the ancients. If a vessel of water is kept fiill by 
. ^^ runniBg through it, and a Me is xnade near tl^ 
bottom^ the water will run out of the hole and fill any 
other vessel at an nnifonn rate; and this other vessel may 
have its sides graduated, or carry a floating index pointing 
at a graduated plate divided into hours^ and even minutes^ 
if the index rises fast enough to distinguish them. When 
the index gets to the top it might be contrived so as to 
open a valve and let the water out^ and itself go to the 
bottom again and shut the valve ; and if this was made 
to take place every twelve hours it would make a self- 
acting clocks requiring no assistance except &om the stream; 
If, on the other hand, a given vessel is filled and gradu- 
ated according to the rate at which the water flows out, 
which is not uniform, but varies as the square root of the 
lieight above the hole at which it stands, this also will 
serve for a clock until the vessel is empty: 

^e anovoamm. 

12; fiand rammg out of 4k. iibfo a^ Hbe b6tMih 
^. a Toad >k)Wfe]* notf irii& vtftry ^etuiy Ittiifti^rtti 
Tebcky; fofr oik' aaxmA ^ 'At fri^tiktt be^o^/flfe 
pwiJckflTi wUMl does- not exiiaA^ » a: Adid, the MuS 
only ialls out, «ad' is nol {Misdpl d8it> bf VzrtiJM! of 
tlie flame pnnfl9pi& wUeh e&able» «t^d to-n^ uf ft ^be^ 
blasrtaitg ]iole^» Anautrt^ctf ^p^nsofe <^«in pisli i 
considerable thickness of. MOid ilanyogli ft tab^ ^itbotrt 
burstmgit. I have seen no account of an index fixed to 
•im bdinvglMd^ that is to a s^nd-gkssr; Ircct <m6 might be 
Blade dQ fiie plaen c^ those tojv in* ifhich isand runs into 
a smafl budiat^ which tiEms ever when it is ftfll^ moving 
BOBie figoiie df ft Mm ^ aninud wi1%.' ft. Hds might be 
made ta t&lsef jfece e^ eirery imnttte or other intervial^ and 
might issm^ tr dock-hani accordingly. Since 1;he invention 
^clodsB idefMsydras have been madd as trtiriosHies^ probably 
much mows perfect <^n any that existed before. Bat these 
things now bel(mg lathei! to the department erf * philosophy 
in sfort/ tihtan ^ sdence in earnest/ and a description of 
•Saaok woold be oat of place in a practical treatise on clock- 

.'" -i.'-'l f \:; 

f . . « * T < 


'\ ' ' r J .••■ - ■', • ' •..■•••.■ . .•.-.'. ', 


-ON ... 

'.■■''.'.• -' ' .' ■ ' • • ■ ' 


j ".:'■■% ; •'• ' ■ ' ' . ■ ■'• ' • 


. • i ..»••, t"*^' ^ « ^ <t tii >^ i t *t»i i j i> 4« n -^ "^ly* ■■^i"- 

i. •■••.'•■ . - 

. ...-,. ... .'CHAjrrwJ-. 

ISr A <3l^ck baa i^ometitnoa Inte dfefineft to Le a 2Ki£k 
.0luiie£ot.€0m#f^ibe vlbtatimts of apcndnhun^ aiid soitie ; 
l^ut thk.ia hikr^y.a veoHTeet dbfimticm ^ for il iu^Iies fostly^ 
that tbeclopk faiuijaaihii^todDbirt toooiuittihe vibrstiond^ 
wbeoa^ ii baa akK) to Biaiataiii them, a^. & pendolum witt 
iioi gD 0B:l(mg endnging bj iteelf ; istca&dlj} tbat the «lodi^ 
fices not alter.tbe tiBieofTibratiQii: ^vrhkb it does* find tliiidly^ 
thattpen^yteiaflr aDii» o*he» vib«tmg bodjrw essential 
toa elock : 'which il is not ; forolocksne aetudljn^ide on 
the pciiaoipii^ cf hismi^ the velocity of ihieir moinement^ 
detenaiaed by a. revolving fly, either with fans to which the 
air offers a Kcsistiaioe^ m wiih a heavy xim^ whose own 
mametU <f inertia prevents it from moving tiw rapidly* 
A more cotreot defiiutian would b^ that a doek or a watcH 
is a machine, ^(maktiiig of a trakt of wheeb turned by a 
vvvighty a fspmigy or any other nearly resistant fisrce^ and ti 


which the velocity is regulated bj attaching to it a pen- 
dnluin^ balance^ or fly wheel, which always vibrates or re- 
volves nearly in the same time. And the only distinction 
(except the arbitrary one of mere size) between a clock and 
a watch is, that a watch will go in any position, bnt a dock 
only in one. 

14, The invention of clocks with wheels is ascribed to 
Pacificns, Archdeacon of Terona, in the ninth century. 
Clocks (without water) are said to have been set up in 
churches towards the end of the twelfth century; and there 
is a story of a clock being erected in Westminster Hall 
in 1298, out of a fine levied on a lord chief justice ; and 
near the same time a dock is said to have been put up 
in Canterbury Cathedral, and one in Wells Cathedral in 
1825. Mention is also made of a clock, apparently of 
some new construction, invented by Bobert Wallingford, 
Abbot of St. Albans in 1326, and which was going in 
Henry the Vlllth's time. Prom these and other notices it 
seems pretty clear, that, though the earliest clock of which 
the actual construction happens to have been preserved, was 
that made by Henry de Wick for Charles the Vth, in 1379, 
yet he is not to be looked upon as the inventor of them. 
According to the description given of that clock, it differed 
in nothing, except in having a horizontal balance iostead of 
a pendulum, from many old church clocks still in existence, 
being merely a thirty hour dock with one hand; and the 
striking part was exactly the same as is still used : in fact 
in some respects it exhibited a more advanced state of me- 
chanical art than the clock (I do not know of what date) not 
only in existence, but in action, in Peterborough Cathedral ; 


wMch has a wooden frame instead of an iron one^ and 
instead of being wound up by a key or winch, is wound up 
by long handles or spikes stuck into the barrels. It has 
bowever a pendulum. The going part of the clock has 
indeed lately been superseded by a modem one; which 
is far less creditable to the mechanical skill of the time at 
which it was made than the old one, especially consideriag 
that it has no dial to work, a circumstance which affords 
unusual faciUties for a good clock. The old striking part 
still does the striking on a beU of considerable size, 

15. !From these old church clocks have descended all 
the modem race of smaller clocks and watches, which have 
arrived at a degree of perfection which seems truly won- 
derfol, when it is considered that, though there is no 
such thing in nature as a perfectly isochronous pendulum 
(one which vibrates different arcs in the same time), and no 
such thing as a train of wheels with perfectly uniform action, 
yet penduluias can be kept vibrating with no greater devia*. 
tion from isochronism than one beat in half a million. In 
the mean time the church clocks themselves have descended, 
. in the hands of all but a few makers, into little better than 
ironmongery : and many of them display the grossest ig- 
norance, not only of horological, but of the commonest 
mechanical principles. Perhaps the most striking instance 
of neglect of horological principles is the practice, of whidi 
Mr. Vulliamy in his ' Considerations on Public Clocks' 
gives several rastances, of putting fans or wings to the 
pendulum; I sappose, for the purpose of preventing it 
from occasionally swinging so far as to drive the pallets 
into the scape wheel under the iofluence of such a weight 



as was found necessary to cany tbe tram through all the 
occasional impediments arising from bad cutting of the 
wheels^ dirt^ the force of the wind upon the hands, and all 
kinds of mechanical defects. It is remarkable that, until 
lately, the French have been much in advance of us in this 
largest kind of horological engineering, and have spent 
much larger sums upon their public clocks. Mr. Yulliamy 
mentions no less than four in Paris, which appear each to 
have cost about £1000, .exclusive of* some other expensive 
appendages, such as enamelled diab, and the bells. There 
is not a clock in England which has cost anything near 
that sum, exclusive of chimes and other appendages, which 
do not strictly belong to the clock. The estimates for the 
Great Clock for the Jfew Palace at Westminster indeed 
exceed that amount; but that is to be a perfectly unique 
specimen, combining unusual size and unusual provisions 
to secure accuracy of 'performance,' as the clock making 
phrase is. 

Of late, however, an improved style of turret-ckx^ 
making has been introduced by some of the best makers. 
And since the erection of the clock at the Boyal Exchange, 
which contains several contrivanees never before used, there 
can be no doubt that, if we choose to pay toi it — and leaa 
than half the cost of the above-mentioned French clocks 
is sufficient for all but the largest clocluH-we can have large 
elocks made just as well as small ones. The present as- 
tronomer royal (who certainly cannot be accused of undue 
partiality to his own countrymen in matters of science) has 
said that he has no doubt that the Exchai^ dock is 'the 
best public dock- in the world/ and thrt he f believes it is 


BapericMT to most ftakonomioal clods in tlie steadiness of its 
tate.'^ Of course this degree of accuracy cannot be ex«> 
]pected in imy but &e v^ best docks; but by merely 
attim&ig to a hw simple considerations they may certainly 
he ioade, and irithoBt any great addition to their cost^ to 
•pproooh to the accuracy of all but tiie most highly finished 
astronomical docks or Tegulators, instead of being, as thef 
frequently are^ merely the public exponents of the time of the 
best clock which the man who 'looks afl;er them^ happens 
to possess, or of the nearest railway station clock, which is 
probably a small spring-clock tiiat cost £6, and is set 
right when it gets as much as five minutes wrong. . 

16. It is evident that a fly, either witii fans or weights, 
would, if the resistance of the air and the friction of the 
maohinery w^ie uniform, produce a constant velocity in a 
train of wheels turned by a weight hanging by a string 
unwound from the axis of the first wheel in the train. One 
would imagiBe therefore that tiiis would be the earliest form 
of a dock, as it is the most simple, and a fan-fly was 
actualfy used in De Wick's dock, as it still is, to regulate 
the ¥ek>cily of the sinking part. Whether this was so or 
not, we haw no meansof knowing; but it is worth noticing 
ihat that oonstnidion, or a slight modification of it, is 
actually in toe now. The dock which is nsed to keep the 
great «qaatoreal tdescope at Cambridge in motion opposite 
to the motionof the earth, so that it may remain for some 
time pointed to a given dtar, and not move by jumps as it 
would if dm«n by a dock with a vibpitmg pendulum or 

# * ' 

* Parliamentary Papers respecting the Great Clock for the New 



balance^ is r^ulated by a governor j like that of a steam 
engine; which is also called a conical pendulum^ because 
the rod of each ball describes a cone as it revolves. As 
the force of the train increases the velocity of the balls 
they fly farther out^ which increases their moment of 
inertia or resistance to motion^ and may also be made to 
apply an increasing Motion by means of a spring to some 
part of the clock, and so restrain the increasing velodtyi 
as the governor of a steam engine regulates the supply of 
steam to the cylinder. Such pendulums have not yet been 
made to approach the accuracy of a vibrating pendulum 
with an escapement, being much more affected by any 
variation in the force of the clock. The time of one 
revolution of such a pendulum in a circle (which it cannot 
always be persuaded to describe) is the same as that of a 
double oscillation of a vibrating pendulum whose length is 
equal to the height of the point of suspension of the 
revolving pendulum above the plane of the circle in which 
the ball revolves; which will of course vary as the force 
that drives the pendulum varies. There are other practical 
difficulties in their use; ajid it ai^>ears to me that uniform 
continuous motion of a train that has heavy and therefore 
variable work to do is more likely to be obtained by the 
method which I shall describe under the head of train 
remontoi/ren (176), than by any contrivance for reguUting 
the motion of a conical pendtilum by friction. 

17. Most people know that a clock consists of a train 
ef wheels, generally four, of which the lowest, if it is an 
eight day dock, turns round in about twelve hours, or 
requires fourteen turns to wind it up for the w^, m3i the 



tt^hest tnnia in a mmate. l%e two intermediate wheeU 
ate msteij required to cany off the difference of velocity 
between the two extreme ones and to work the hands. 
Tbe lower of ^ese two intermediate wheds is osnally mad* 
to turn once in a hoar, so that the long or minnte hand 
may be set ajxin its am or arbor produced as far as the 
face of the clock ; and the second wheel has nothiag to do 
hut to reduce the multiplier of Biity, or the ratio of the 
velocity of the highest to that of the entire wieel, as the 
one which carries the long hand is usually caUed. It may 
he convenient to give a sketch of the common arrangement 
of the going part of a dock, with the different parts marked, 
boia which their action will be pretty evident. 

18. The string is coiled 
aixteffli times round the 
barrel when the dock is 
wound up. The barrd asis 
or arbor has s sqnare end 
to fit the winding k^, and 
when yon turn it to the 
right the iat«hett teethe of 
which three or four are 
shown in tJie drawing, raise 
the dick which is fixed to 
the great whed and kept 
pressing gainst the latch- 
ett by a spring; but when 
yon stop winding, theratch* 
ett teeth cannot pass the 
other way, and so the weight acts upon the whed, tending 

19 cxioccs. 

t6 Htmnf it'to the left ' The.^ieat wheel has imOSlf 
^inel^Hrix tKA> andfitt dbdvee tbeditile wh^ (>r jpifiidii 
of jldul QBtttrft wlKe^ lAkhi hae cigbi teeth ^* kim^ aM 
ilMnfq^e goe^: Wneltt tinies as^ 1^. ^revt wl^* 
inbftfqmtrfa^vHka^ chlvirfttliti {liw^feuel: ti^ej seaond wfa^ 
frliA iAv«a fto mipf iriM^^mA '%mk teetk an^ \smm 

ofieltTOSi oil tte saap^ wliael pinion^ jK^of fli& aecSond^wh^ 
pfiliasiy^kiid i^' ited) i^^ iiid ntaoibaar/of the to^tli thai dma 

tiiem respectively, — ^ — ^aay = ffO, The only conditions to 

lie obeereed are, that no driven pbdon onght to have kaa 
than seven leaves at the lowest (except of a land whidi I 
ahaH mention hereafter) ; and that if the teeth of both 
iseheeb are of the same size,, the centre wheel nmst \m^ 
more than, or at least as many as the aeeond, or they 
couldnotbegotinto the frame togeflvar. Suppo-e eitfw 
jpj or jt?, to have seven leaves aind. tiie othar:eigl|t^. the^l 
i^ t^ mnst = 60 X 56; and those twj> ntambera 6Q i»id'5Q 
inU do for the teeth of those twb. \idiedls #s.W9ll a3 any 
jt)ther numbers giving the same pto^^ct. . . % 

19. With regard to the weight, the t&s^t <will ob^csrvf 
that it hangs by a double line, or' a siDgle moveabid piiillej^ 
the effect of which is that only half the weight realfy acta 
iipon the clock; but then the string is ilpon the whafo 
flouble the length of the actual- fall (tf the wei^t, aM 
therefore the effect of the weight is jog^. the -same if it Wera 
hun^ by one string and had the same fSedl; in which caao 
the barrel would of course have to be only half the size, or, 
the. number of teeth in the wheel doubled, or the nurnb^ 


of lea?es in the centre wheel piiiiim ' haired. Li com- 
padiflons of clocks one aometmies sees calculations of what 
the weight is eqniyalenbto 'on the single Une/ or what 
IB worse^ a mere statement of the weight withont any in* 
Satmation as to its taH, which is jnst as necessaiy an 
ingredient in the comparison. It does not signify (oLcept 
as regards the friction of the pulleys, which cannot be re^ 
dnoed to calcaktion) what the weight on the single line is, 
^r whethip there is one line or a dozen : a given weight 
fidKnga'gi*«n Knight is equivalent to itaelf, andtonothing 
^se, except anothv wioghb falling a height as much greater 
« less as flifi weight is lessor greater. Therefore in every 
possible .case if jon want to see how much of the ' first 
power' of the dock is wasted in the Motion of the train 
andpuDfiy^ or.howmTidi.a;givmycap«nBBtreqniieg, you 
have nothing to do but^ to \ remove the .pallets and fix to 
the scape wheel or its arbor an arm of any lensth that is 
conv Jent, and hang small weights to the end of it, till 
y<m finid what w^ht the dock will just deddedly'lift, 
Hm atm being horizontal. The arm should have a cpun« 
terpoise and be as light as possible^ so that its own weight 
may not enter into the question or add much to the Mdion 
of the iSttape whed pivots. Now if you know the fall of 
the <doak. weight for eight days, you know it for one day, 
and one minute, or one revolution of the scape whed, 
whatever it may be, or any aliquot part of that revolution; 
and if you know the length of your measuring rod you 
know that its end would move through a space 6*28 
times the length of the rod in one revolution : or, without 
that calculation, you may turn the scape wheel through 


^ne-eigHth or one-tentli of a revolution, and measure liov 
much the little weight rises therein. Sappose this rise is 
6, and the corresponding fall of the great weight a; 
then the great weight w ought to be to the little weight m 
as b: a* It will in fact always be a great deal more, and 
the amount of the excess oiwa over mbin different clocks 
shows the comparative loss of power in their trains. 

20. And again for the converse of the experiment aa 
regards the escapement (of which I shall treat presently) 
when you know what the momentum of the clock weight 
really amounts to by the time it reaches the scape wheels 
you know how much of it is consmned in driving the 
train, and how much goes to drive the pendulum. These 
proportions will be found very different in different clocks. 
But it is time we should proceed to exphiin what that im. 
portant part of a clock called the escapement is. 


21. An Escapement is a contrivance by which the 
teeth of a wheel, commonly called the scape wheel, are 
successively let go and stopped at certain short intervals of 
time. In all the common kinds of escapements there are 
two pieces of steel called pallets fixed to an axis which 
vibrates, so that as one pallet moves out of the way of the 
teeth, the other moves into the space between some two 
teeth, and so stops j;he wheeL Consequently as every tooth 
has, at some point in the revolution of the wheel to dear 
both pallets, there will always be twice as many beats in one 
revolution aa there are teeth ia the wheel. But the es- 
capement must do something else : the pallets must be of 


flach a sliape that as the teeth escape tliey shall give a posh 
to the pallets, and this posh is commaiiicated to the vi- 
brating axis, and to the peiiduliuu or balance, which is 
attached to it. A baltmce is simply a &y wheel, and was 
no doubt first called a bahmce, because in the earliest clocks 
it was in the form of a balance and not of a wheel, con- 
sisting of two arms set npon a vertical a^ and canjdng 
weights hiiog at the ends of the arms. This was the form 
(rf 1i,e balance of De Wick's chick : aS^^wards the weights, 
instead of being hung on to the arms, were screwed on, so 
that their distance from the axis could be adjnsted more 
accurately. The escapement was. exactly the same as that of 
a bottle- jack, or the commonest kind of watch, and is called 
a verHcai escapement. The pallets are two flat pieces of 


f^j&Ap,p, set on to a vertical axis^ in planes aboat at rij^ 
angles to each^other. In this drawing the highest tooth (rf 
the wheel (which is from its shape called a crown wheel) 
fe represented as just escaping and coming forward (or 
towards the reader) ; the lowest tooth then strikes against 
the lower pallet^ which stops it; and not only stc^ it^ 
but, as the vibtation of the balance in that direction cannot 
be suddenly stopped, the pallet advances a little further 
forwards, and so brings the wheel back a little, and produces 
what is called the recoil, which is, in a less degree, com- 
municated to the rest of the wheels, and ultimately to the 
weight which drives them. It is evident that each tootii 
as it escapes gives a push or impulse to the pallet wUicb 
presents a sloping face to the direction of the tooth; and 
so the time of a vibration in this clock d^iended upon the 
force of the impulse relatively to the moment of inertia of 
the wheel, except that a greater force would catise it to 
make larger vibrations. The effect of such an- escape- 
ment as this may be judged of by taking the penonlum off 
a common clock, leaving the cratch on, and observing the 
difference between the time the hands will take to make a 
revolution when the weight is in its original state and 
when it is added to. In a bottle jack the piece of meat 
and the iron wheel which is generally hung to the jack 
together form the balance. The jack is driven by a spring 
instead of a weight, and the velocity of rotation of iihe joint 
diminishes as the [Spring runs dpwn. The ^cape wheel of 
the vertical escapement must either, have an odd number 
of teeth, or the axis of the pjJlets m^st be a little on one 
side of the centre of the wheel. 



22. Clockft remained with balances it seems for about 
300 years after De Wick's time; And there was a great 
dispute as to the first inventor of clock pendulums. It is 
usually stated that the fEumous Galileo discovered (as he 
3uppo8ed) £rom observing the swinging of a lamp hung 
firom the top of a churchy that pendulums oscillate 
through different arcs in the same time: a property 
which is called the isochrpnism of , the pendulum. It 
is said, however, in the able and 'interesting artidd on 
clocks in the Ewsj/clopadia Britannic^, (last edition), that 
this cannot be reckoned among the ^S|fH)veries of GaU* 
leo, ' for the andent astronomers of the East employed pen-* 
dulums in measuring the times of their observations, 
patiently counting their vibrations during the phases of an 
eclipse or the transit of the stars, and renewing them by a 
little ^lih with the finger when they languished : and Gas- 
sendi, BicdoU, and others in more rec^t times followed 
their example.' 

It is iLot indeed strictly true that pendulums do vibrate 
different arcs in the same time ; but it is true, that so long 
as the arcs do not exceed a few degrees, the oscillations ar& 
very neady isochronous. And, therefore, if a pendulum can 
be kept twinging nearly the same small arc, its oscillations 
will be so nearly isochronous, that the difference will 
not be sensible except in a considerable time. This dis- 
covery would naturally lead to the application of the pendu- 
lum to docks : and as is usual in such cases, the invention 
was probably made independentiy by several persons about 



the same tinier And in favour of the daim of Harris^ the 
clockmaker^ who is said to have made a pendulum clock for 
St. FauFs^ Covent Garden^ in 1621^ several years before 
Dr. Hooke^ Huygens^ or GraUleo's son^ who all claimed the 
inve^tion^ it may be remarked^ that if, either by accident or 
design, he placed a clock with the axis of the balance hori- 
zontal instead of vertical, and left one of the arms without 
its weight, he wonld see that he had made a clock with a 
pendulum, of which the isochronism was probably by that 
time generally known, and he would naturally adopt it 

23. But whoever was the inventor of pendulum clocks^ 
there is no doubt that Huygens was the discoverer of the 
true theory of the pendulum ; and though his application of 
the theory to practice is now abandoned, yet as aU pendu- 
lum calculations depend upon it, it is proper to give a short 
explanation of it. He discovemed that the curve in which a 
body must move so as to oscillate large and small arcs in 
the same time is not a circle but a cycloid; which is 

the curve genera- 
ted by a point P 
in the circumfe-^ 
rence of a circle 
D E P rolling on 
® a straight lineB CL 
And he also found 
that it has this re- 
markable proper- 
ty : that if the cy- 
cloid B P P C ia 


cat in two at its lowest point "E, and the two halves put 
together as in this figure^ CF being placed in the position 
^'B, and BF in the position AC^ then the end of a string 
A P of the same length as AB will^ as it unwinds firom A B 
and winds on to A G^ redescribe the original cycloid B P F C^ 
This is expressed mathematicall j by saying that both the in» 
volute and the evolnte of a cycloid is an equal cycloid. Con* 
sequently if a weight or bob were hung on to the end of the 
string AP^ it would be a pendulum swinging between the 
cycloidal cAeeis A B, A C, and describing the cycloid B P F G; 
and such a pendulum would make large and small osciUa* 
tions in the same time^ or would be perfectly isochronous. 
Nevertheless it is found that the impossibility of making the 
cheeks so accurately as to cause the pendulum to vibrate in 
a true cycloid^ as well as other causes, all apparently insigni- 
ficant, so much disturbs the isochronism of such a pendulum^ 
that it is more isochronous, even up to arcs of 6*^ or 7^, with- 
out the cycloidal cheeks than with them. 

24. It will now be evident why pendulums swinging in a 
circle in small arcs are nearly isochronous, and more so the 
smaller the arcs are; for if you describe a circle with the 
radius AF, it will very nearly coincide with the cycloid for 
a short distance near F, the lowest point of the cycloid, but 
will deviate farther from it the farther you go from. F ; but 
the smallest arc in the circle really takes more time to 
oscillate than the largest in the cycloid. The difference 
between the time of any small arc of the circle and that 
of any arc of the cycloid varies' nearly as the square of the 
circular arc, or of the angle which it represents; and again 
the difference between the times of any two small and nearly 

88 CLOcfcs. 

equal circular arcs vanes as the arc itself; tliat is^ if a cloek 
pendultim is swinging 4^ from zero^ and from some cause 
the arc is diminished 5' or 10'^ tlie clock will gain twice as 
mucli as if it liad only been swinging 2^ when the same 
diminution took place : which is just the opposite of what 
would probably have been guessed. This is called the 
circular error, and if a is the arc, and da the decrease of it, 
the gain of the clock in the day (supposing it to have a 
seconds pendulum) is 10800 ada sss rather more than 1 
second, if- « = 2° or 'OSS (the radius being unity) and 
da 3= 10'. 

26. It is well known that a two seconds pendulum is 
four times as long as a one second pendulum : that is to 
say, the time of vibration of a pendulum varies as the 
^square root of its length. But this does not teOi us what 
the actual length of a seconds pendulum must be. Let 
; be this length, and t the time of one oscillation measured 
in seconds ; ir the ratio of the circumference of a circle 
to its diameter, or 8 '1416 ; ff the force of gravity, which 
is expressed by the number of feet per second, or the 
velocity with which a body is falling at the end of one 
second, or twice the height that it faUs in a second ; then 

^stty/— . Ill England ^=32*2 feet; and therefore the 
length of a simple pendulum to vibrate in one second of 
mean time must be 39* 14 inches nearly. And in like 
manner the pendulum to vibrate sidereal seconds must be 
39'14x997*=38-87 inches (§7). I may mention here 
that a clock with a pendulum for mean time may be made 
to show sidereal time (though it will not mark sidereal 
seconds) on another dial, by the ^addition of four wheels 


and two iarbors as follows. Put a wheel oi 247 teeth ou 
the centre-wheel arboi" of the clock, and let it drire a« 
wheel of S31 teeth on an arbor which also catries a whedk 
of 48 ; and if that drives one of 32, this .last wheel 
willserve for the ceatre wheel to carry the nanate hand 
of the sidereal dial; .as it will torn slower than the othep 
oentrewwheel in the ratio of sidereal to mean time within 
a fraction of no more than -nmr of a second in a day."^ 

I spoke jnst now of a eimple pendntum. A simple 
^ndnhm is, what can only exist in theory, viz., one 
in whieh the rod has no weight and the bob consists 
only of a singly heavy point. The weight of this point 
or oob is of no consequence in the theory of the simple 
pendulum, though it is, as we shall see, of great con- 
sequence in practice in resisting the various disturbances 
to which a pendulumi is subject. Every real pendulum 
however (which in mathematics is called a compound pen* 
dulum) is equivalent to some imaginary simple pendulum; 
and the^thirmer the rod is, and the heavier the bob, the 
more nearly the real pendulum will approach in length to 
a sunple one of the same length, measured from the point 
of suspension to the centre of the bob. When the rod 
and th6 bob'are of any regular shape, as they generally are 
in clocks, it is not very difficult to find by calculation the 
length of the equivalent simple pendulum, or the distance 
from the axis of suspension to a point in the real pen* 
dulum which would be the bob of the simple pendulum if 
the whole mass were collected there. This point is called 

* These immbers are taken from the Philosophical Magazine for 
Peb. 1850. . 



the eetUre of ascillatum. But it is not like the centre 
of gravity, a fixed and independent point, for there is a 
different centre of osdllation for every different centre of 
suspension. It cannot be explaraed how the centre of 
oscillation is found without reference to another imaginary 
quantity called the radi/us of gyration. If M is the whole 
mass of a body revolving or oscillating about any axis, 
and i ia the radius of gyration about that axis, then Mi|= 
the sum of all the elements m of the body each multiplied 
into the square of its distance r firom the axis, which may be 

expressed as 8 (*»/•*) ; and therefore i * = m '^^ 

som can only be found by means of the integral calculus 
in every particular case. But there is this curious property 
in revolving bodies ; that if we have once found the radius 
of gyration k^ belonging to an axis through the centre of 
gravity, then the It belonging to any parallel axis at the 
distance A from the former one is given by this equation 
i*=^J-i-A'. Suppose then that we know what ii is 
for a body of any shape of which a vertical section in a 

plane perpendicular to 
the axis of suspension is 
A B C D revolving round 
its centre of gravity G, 
and we make it revolve 
or oscillate about a new 
axis through S, and we 
want to know what is 
the length SO of the 
equivalent simple pen- 
dulum ; it may be proved 


that SO = ii+|?! and /. GO = SO — SG = |1. 

Prom this we may see also that if we draw two circles 
round G, one with the radius S G and the other with radius 
O G, we may put the axis of suspension anywhere in either of 
the two circles and still the body will oscillate in the same 
time. Perhaps therefore the radiu8 of oscillation would 
be a more correct term than the centre of oscillation* 

Now since S G = ^ it is evident that the nearer the axis 

of suspension is to the centre of gravity^ the farther the 
corresponding centre of oscillation will be; and this is 
the reason why a small and delicate scale-beam will 
oscillate as slowly as a pendulum many feet long; and it 
also suggests a mode in which a pendulum might be mads 
to oscillate veiy slowly in a small compass. If we make a 
pendulum with two bobs^ one above and one below the 
axis of suspension^ or take a wheel with a heavy rim and 
suspend it on an axis a Uttle out of its centre, it may 
vibrate two or more seconds in no greater space than a one 
second pendulum; but such a pendulum will not be so 
effectual in resisting the disturbances of the escapement as 
a long pendulum of the same weighty which is what we 
want a pendulum to do. 

In all clock pendulums the effect of the weight of the 
rod is to throw the centre of oscillation a little above the 
centre of gravity of the bob, though below l;hat of the 
whole pendulum; and by a few trials of the bob at different 
heights the pendulum can be made to vibrate in the proper 
time without the trouble of actual calculation. Now on 
looking at the drawing of the cycloid, it is easy to see that 

42^ . CLOCKS. 

the ,o&abn 6t tnepeSimoii of anresd {i.e. not a s&ii|)ld) ^cxh 
dulum swinging between cjdoidal cheeks is contiiraally 
changing, and therefore the radios of oscillation is- con- 
tinuaUy changiBg, and deviating from the length reqnired[ 
to describe the cycloidal arc. In other wordi^ a real pen^ 
dulum swinging in cycloidal cheeks is not equivalent to ^ 
rimple pendulum swinging in a cycloid, and therefoj* ii 
not isochronous. 

26. Pendulums are now almost always suspended by 

then swu^ without any friction; and they ate connected 
with the axis or arbor carrying the pallets by a fork ot 
ertUch, which is a light rod or ' aarm coming down bom th« 
pallet arboif as if it was intended to'^be- the pendulum^ 
but ending in a fork whidh embrstc^ the penduhutf 
tod ciosdy btit not so tightfy asi to jHre^ent the roci 
from sliding wifliin itj so that thlfe pendulum and th^ 
crutch move tbgether, aAd the palfet arbor vibratei 
as if the pendulum were hung ' directly to it in thd 
iold way. It is to be observed; that as the spring does not 
admit of being bent suddenly like a string, but assumed 
a curved form as the pendtdum swings, its effect approaches 
to that of the cycloidal cheeks. It has accordingly been 
^attempted to make the spring of such a form and strength 
as to render the vibrations of the pendulum isochronous ; 
but without success; and it is of much more consequence 
to find tiie spring which requires the least possible maintain- 
ing force to keep up a given amount of vibration of the 


^iaid«lttm^ "^ieli is pidlubUy as tbixi a «pnp|^ as » stffefov^ 
&& weight' of tha petodoliun* t 

f I am aii^ thdt smw^ experiments have bee^ 
tad lefaited to in the artide on dock^maUng, among iher 
f nwrnnftiititafes/ in the JSny^lopadia MetropoUUma, hosni 
whiehit would appear that a spring -008 of an inch thiol: af* 
Sected tlie vibiatidns of a pendalum of 141hs.iireight less than) 
any other spring of the same length and vidth, either thidcer 
or thinner. I must say I should have had some difficulty m 
believii^ that there was not some mistake in such a result^ 
from the thinner springs having got what is called a set in 
fixing them (which is very likdy to happen with very thin 
ones) or some other cause^ even if I had not been told also 
that similar experiments have been made by other personswitb 
the diffisrent and more probable result,* that the thinner th0 
spriog) without limit, the less it ^ects the peodulum. And 
as spinigs canno^ be nsed with safety of sndi thinness 
hk paoportionrtb ^ weight of the pendulum. as that abova 
qientionady we may in any case adopt theplacticalcdndosiom 
that the thinneaf the spring is the less it will affect the peUti 
dnliim/tUitiiiE^ tike less itarate wiU differ from a pendtilum 
JMUgmg fiom an^aads of infinitdy small thickness* 
i 27« It is of great importance that the real point ci 
aaspenaiQn Df the peiidulum, that is the top of the spring 
where it b^ins to bend, should be kept firmly in the same 
place ; for if it moves it will increase the time of vibratiaDy 
since this is evidently the same thing as if the fixed or real 
point of suspension was a little higher up, or the pendului^ 
80 much longer. For this reason, in tl^ best docks, the 
cock which carries the pendulum is a strong piece of brass. 


or in large clocks a cast-iron firame^ firmly fixed to the wall 
at the back of the clock. In order that the pendulum may 
hang so that the spring will have no tendency to twist it as 
it swings^ the top of the spring is pinched or 'dipped' 
between two thick pieces of brass or iron called cAop^, and 
firmly screwed there; and these chops have sqnare ends 
exactly at right angles to a line down the middle of the 
spring. A little way below the top of the chops, and 
exactly in the middle, a strong steel pin is put through 
them and the spring between them, at right angles to the 
plane of the spring, and this pin has shoulders, so that 
ita thin ends beyond the shoulders wiU jnst drop into two 
nicks or Vs in the sides of the cock with the shoulders 
resting against the sides. It is evident that the effect of 
this will be that the weight of the pendulum will cause the 
square ends of the chops, and therefore the top of the 
spring, to be horizontal; and so, if the pendulum is made 
symmetrically, as of course it ought to be, it will vibrate in 
a vertical piano- at right angles to the line which is the top 
of the spring, without any tendency to twist. 

28. 'The two Vs should be made as nearly level as 
possible, and the clock frame must be so placed that the 
pallet arbor is exactly at right angles to the plane^of motion 
of the pendulum. It will be easily seen if it is not, because 
then the pendulum will sHde backwards and forwards in the 
fork by which it is connected with the pallets. In the 
chapter on church clocks I have given a drawing which will 
show the suspension of a pendulum* In common clocks, 
both house and turret clocks, the cock is fixed to the dock 
frame, and has merdy a slit in it into which the spring fits. 


having a piece of brass rivetted on to the top to keep it 
from dropping through the slit. And this slit is very often 
made so crooked^ or oblique to the axis of the pallets^ that 
there is considerable friction of the pendulum sliding in the 
fork; and the cock is generally made so slight^ and the 
dock itself so loosely fixed to the clock case^ that the motion 
of the pendulum may be plainly felt if you put your finger 
on the clock. 


29. Occasional mention may be seen in books of 
pendulums vibrating for several hours on haife edges ^ that is 
on a suspension much like that of a scale beam. Such 
a suspension is the best for experiments to ascertam the' 
undisturbed time of vibration due to gravity only, excluding 
the elasticity of a spring] but it is not to be inferred that 
it will answer in practice for a permanent pendulum of the 
proper weight for a clock ; for even if the knife edges and 
the planes on which they stand are made of the hardest 
stones^ it is foxmd that they soon suffer from the severe 
pressure, and introduce an amount of friction which is fatal 
to the accuracy of the pendulum. 


SO. A suspension on friction wheels has also been tried, 
but generally without success, the pendulum pivots always 
wearing a hollow in the wheels; and then of course they 
ceased to roU, and produced an amount of friction which 
was fatal to the proper action of the pendulum, Mr. Y13I- 


liamy, howerer, st&tea in the papers re^eoting the West- 
minster dock, that the pendnhim of the dock at tlie ¥oBt* 
office, which wdghs 4 cwt., has been goii^ erer sinee it wu 
put i^ apon friction roOeaia ; that is to saj, npon so nmch 
of an entire fdctioa vhed as is required for t^ small angle 
thiongh which the pendulum vibrates. The arrangcmmt is 
shewn in this drawings 
The bearing feces of 
these portions of fric- 
tion wheds are not in* 
deed made sensibljr cir- 
cular, but are Sat pieces 
of vay hard steel, with 
certain provisions for se- 
curing to tbe pivots a 
flat bearing whidi it is 
not neceasary to dra- 
cribe ; but as tiie radius 
of the pivots is only 
l-150th of that of Ha 
friction wheds, which 
9&. 6in. Itmg, imd 
the rollers only move 
through an angle less 
than 1' on each side ot zero, the versed sine or measure 
of the curvature of a friction wheel for that angle is too 
small a quantity to be expressed by a table of logarithms of 
seven figures ; and th^efore {at all practical purposes the 
bearings are the same as if Ihty were t^lindrical portions 
-^ entir« &iotion wheels. 



Mr. Yulliamy tells me that bom the weight having 
^been thrown almost entirely upon the pair of rollers nearest 
to the pendnhim^ and the faces having been allowed to get 
idirty^ it has been necessary to repolish them once lately. I 
shonld be inclined however to make the pivots rather 
thicker than *4 inch diameter for a Yerj heavy pendnlnm; 
und the rollers should be placed nearer together^ becanse 
the farther they are off the greater the pressure is upon 
them, in ihe inverse ratio of the cosine of half the angle 
included between them. It happens that for 60^, the angle 
at which they stand in the Post-office dock, they have to 
.bear just l-4th more pressure than the actual weight of 
the pendulum. They would act at any angle safely above 
that to which the pendulum swings : 20^ between them 
wou]ld give them a sufficiently wide stride for a firm bearing 
4nd add hardly anything to the pressure, and would also 
reduce the tendency which friction wheels always have to 
twist the pivots. The pivots here are not indeed common 
.projecting pivots, which would be too weak, but a hard 
^iece of turned steel .bedded into a strong beam, to which 
the pendulum is hung, and to another part of which the 
.pallets ace fixed. 

The advantages of this method are, that there is no 
necessity for any ccnnpensation of the spring ; which how-* 
ever, as will be explained under 'compensation of pendu- 
lums,' requires only a small addition (though of rather unoer- 
.tain amount) to the ordinary compensation required for the 
.expansion of the rod. The risk of a spring breaking is too 
,jrare an accident to be worth enumerating as an advantage of 
jUus snspensioi^; and it can be repaired in an hour when it 


does happeiL The principal advantage of it probably 
is that it enables the pallets to be put on the arbor 
which carries the pendnliun, or in other words dispenses 
with a crutch, and therefore prevents the loss of force 
which must always (though to a small extent) take place 
in the transmission of the impulse through the crutch, both 
by the friction and the shake in its pivot holes ; and we 
shall see that everything that dbninishes this force, and 
therefore requires a stronger pressure on the pallets, m* 
creases the chances of error in the going of the clock. 

No doubt even the small amount of rolling Mction 
in this suspension would destroy the motion of a free 
pendulum sooner than a spring: but the proper com- 
parison to make is between a common pendulum with 
a spring and crutch, and this without a crutch. I 
have no means of stating the result of such a com« 
parison, either as to this point or the performance of the 
same clock with the two different pendulums, which I 
should require to be tried before I would undertake to 
recommend such a suspension, Mr. Vulliamy however teUs 
me that the Post-of&ce pendulum was kept vibrating 2** 22', 
with only a weight of 30 lbs. falling at the rate of 47 feet 
in the 8 days (the dial work being unattached) ; which is 
tm unusually small weight and fall for a large clock; and 
I must say (with the reservation just now made) that 
I think this fact is of much more importance towards 
determining the actual merits of the plan than the 
astronomer royal^s inference, in his report upon the plans 
for the Westminster clock, that it would fail because he 
^ knows that it would f^ for a balance or a vertical-force- 


magnetometer:' on which it is very obrioaa to remark 
that knife edges are the only things that imawer for a 
balance, and 7^ they invariably £ul under the continaed 
pressure of a heavy pendnlmn. I need hardly say that this 
mtspenaion most be very expensive, and votdd require the 
greatest care in properly constmcting and adjusting. 

31. Long after pendulums were invented, tiie vertical 
escapement continned te be used, in the fonn in which 
I have Buggeated that Harris may have invented pendulum 
clocks; and snch clocks are still in existence. The arrauge- 
. ment is that shown in this drawing, a second crown wheel 

80 clocks; 

being' used in order to keep the arbors ol all the wheebi 
except the scape wheel horizontal. A wheel of this kind 
is sometimes called a cotUrate wheel; and its teeth and those 
of the pinion which it drives> onght properly to be of the 
sihape required by two conical or bevelled wheels of the re* 
spective sizes of the wheel and pinion. In practice, how* 
ever, the pinion being small is made cjlindrical as nsnal, 
and the teeth of the contrate wheel being thin, they work 
together ^ sufficiently well for the work in which they are 
employed, such as vertical watches. Clocks with this 
escapement however have been quite superseded by the 
invention of ' ^ 


32. These are said to have been invented either by Dr. 
Hooke, one of the most scientific men, and probably the 
greatest inventor of the 17th century, or by Clements, a 
London clockmaker. It will be at once understood from 
this drawing what they are. [See next page, "] The bottle- 
jack or 'vertical' pallets, being close to their axis of vibra- 
tion, required the pendulum to move through a large arc 
in order to clear the teeth of the scape wheel; and besides 
what we have seen of the disadvantages of large arcs, thqr 
require a larger maintaining power than small ones. In 
this drawing a tooth, a is represented as having just escaped 
from the pallet A, and a tooth b on the opposite side of the 
wheel has dropped on to the pallet B. The pendulum will 
not however stop here, but will advance a little further to 
the left, and so the slope of the pallet B will drive the tooth 
h bacj: again a little and produce the recoil, which may be 


observed very plainly 
in any common honse 
clock with a seconds 
hand. The sloped 
&ce of the pallets 
Causes the teeth to 
give th«n an impulse 
in escaping, so as to { 
maintain tte motion ' 
of the pendulum. 
This kind of escape- 
ment is much the 
most common, and 
will probably never be 
superseded, as it is 
sufficiently accofate 
for ordinary purposes, 
,and is very easy to 
make, siiice no p^- 
tictdar form is requi- 
red for the pallets. 

Their acting faces are generally made fiat ; but they are 
better convex, as in the drawing, as ^ere is then less recoil 
and less wetuing of the pallets by the points of the teeth. 
Strange as it seems that brass teeth should wear boles in 
steel made as hard as it can be tempered, it will always be 
found that the teeth have worn boles in these pallets after 
a few years, and the hole will be deepest towards the end of 
the place which the tooth resiches. It is evident that the 
tendency to make this hole will be less if the pallet is cou- 


vex than if it is flat; and accordinglj in the best clocks of 
this construction the pallets are so made ; and care should 
also be taken that they are so placed that the recoil and 
the drop may be equal from each pallet. The recoil escape- 
ment is now abandoned in aU the best clocks, though it 
once had considerable persons as its advocates, who appear 
to have been misled by observing that it keeps the pendulum 
to the same 4irc more than the dead escapement does, and 
thence inferring that it would keep it more nearly to the 
same time; whereas it is now proved, both by experiment 
and calculation, that although, as the force of the dock 
diminishes from increasing Unction or thickening of the oil, 
the arc will diminish, the clock will nevertheless lose ; the 
loss caused by the escapement being more than enough to 
counterbalance the .$gain due to the circular error of the 
decrease of the arc. . The same is the case with a watch 
liaving a vertical or a recoil escapement. 


33. The escapement which is now used in nearly all 
astronomical clocks and in all good turret clocks is called 
the dead beat escapement, and was invented by a clock- 
maker of the name of Graham in the last century. A 
recoil escapement would be converted into a dead escape- 
ment by making the slope of each pallet stop at the 
points A and B where the teeth fall, and making the rest 
of the pallets A D and B E portions of a drcle whose centre 
is C the axis of the pallets. Eor in that case, however far the 
pendulum may swing no recoil can take place. The reason 


vhy this escape- 
ment ia 80 much 
better than the 
retbil escapement 
is, that a variation 
in the force of the 
dock train pro- 
duces hardly any 
effect upon the 
lime of oscillation 
of the pendulum, 
a considerable ef- 
fect upon the ez- 
ieak of its oscil- 

84. This may 
be shown in a ge- 
neral way as fol- 
lows. Let c be 

Uie angle which the pendulum rod makes with tlie 
T^ical o when a tooth begins to act on the sloped &ce 
of the pallet, c' the angle on the other side of the vertical 
at which the impulae ends or the tooth escapes. Theo- 
retically c and c' might be equal, but practically c' must 
be a little lai^er than c, in order that the tooth may not 
drop exactly on the corner of the pallet but just beyond it 
on the circular or dead part. Now while the pendulum is 
descending from c down to o the force of the clock acts 
in the &ame direction as gravity, which is the same thing 

54: CLOCKS. 

«s if the force of gravity, or the earth^s attraction, were 
increased by a certain amount; but while the pendulum 
ascends from o up to c', gravity may be considered as 
diminished by the same amount (assuming the force of the 
clock to be constant throughout) ; and therefore on the 
whole the two disturbances would balance each other, but 
for the fact of c' being rather larger than c, and as the 
tooth evidently makes a larger angle with the pallet towards 
the end of the impulse than at the beginning, the force is 
not quite constant throughout, but a little greater through 
c' than through c ; and therefore on the whole the force 
of gravity may be considered as a little diminished, and 
the pendulum will vibrate a little more slowly than if it 
were free from the clock ; for it is evident that if the 
attraction of the earth were weaker a body would fall, or ^ 
pendulum vibrate, more slowly, as the time of the ascent 
is always equal to that of the descent. It is also evident 
that as the force acts throughout in the same direction 
as the pendulum is going, it must increase the arc of 

S5. Moreover the uniform friction on the dead part of 
the pallets hardly affects the time directly; siace from the 
extremity of the descending arc down to c? it acts contrary 
to gravity, and from c' up to the end of the ascending 
arc it acts with gravity, but as it always retards the pen- 
dulum of course it diminishes the arc. On the other hand, 
in the recoil escapement the action during the middle part 
of the oscillation is the same as in the dead escapement, 
but towards the extremities of the arc the force acts with 
gravity not only in the descent of the pendulum, but also 


ill lis ascent daring the recoil; and therefore my increase 
of the force will make the pendulum swing faster^ though 
at the same time the arc is increased by the action through 
the middle of the oscillation as in the dead escapement^ and 
is not affected by the action at the extremities of the arc. 
The difference between the time of oscillation of a pen- 
dulum attached to a dead escapement and xmattached 
!>eing very small, the second difference, between the times 
)f the same pendulum with different degrees of force in 
tie escapement, must be exceedingly small; which accounts 
for the accurate 'performance' of clocks with this es- 

86. This kind of reasoning however will not enable us 
toietennine the actual amount of the errors arising from 


any particular amount of change in the force, or in the 
arc. The calculations necessary for obtaining these results 
are loo complicated to be introduced here, but the results 
thenselves are sufBlciently simple. Let D be the difference 
in tventy-four hours between the time of the da/s os- 
cillations of a free pendulum and the same number of 
oscilkions of a similar pendulum attached to a clock with 
this esmpement ; a the angle of vibration on each side of 
zero; f the angle at which the impulse begins, c' the 
angle on the other side of zero at which it ends; W the 
clod weight, and A its fall in a day, M the weight of 

the pendulum, and I its length; then it may be proved 

"VTA «'-— c 
that D =T77" Q^ — i very nearly, provided that cf and c are 

not huger than about -r • Suppose for example ^ to be 
2°, ani </- c = 20', W = 2 lbs. and >l = 9 inches (which. 


though less than rumal^ is 6u£5[Gieiit for a highly finished 
clock with light wheels)^ M =s 14. lbs. and ^ = 39 inches ; 
then D wiU = .8 of a second. 

But it must be r^nembered that the daily error in the 
going of the clock is not D, bnt the poritUion of D due to a 
given variation in the force which arrives at the escapement, 
and to the change in the arc of vibration^ which will vary both 
from changes in the force and the friction on the dead part 
of the pallets. Now if the force of the clock is increased by 
a small amount d Vf, and the corresponding change in the 
arc is called da^. then it fdllows> on the principles of thi 
differential calculus^ that the increase of D, or 

To this must be added the circular error, which va 
found to be (assuming it to be entirely uncorrected by tie 
pendulum spring) 10800 o/fo. Therefore the whole daily 
loss of time will be 

,T~r -^ — a 1^;;;^ l + 10800 ada. 

If we took the friction on the dead pefft of the pdlets 
into account, we should introduce anoth^ small term 
showing a farther increase of time, and depending /upon 
the pressure on the pallets and the coefficient off'hcUon, 
or the ratio of the friction to the pressure (which is idd to 
be from I to -jiy between weU polished and oiled siirftces 
of brass and steel), and also depending like the other/ on 
d —c. Both for this reason, and because the dead fri<tion 
diminishes the arc of vibration, the effect of which weBhall 
see presently, it is by no means to be taken for gifted 
that, because a constant friction on the pendulum through 


equal angles on each side of zero produces no direet effect 
upon the time of a vibration, therefore the friction in the 
dead escapement on the circular part of the pallets does no 
hann. We shall see a remarkable proof of the extent 
to which it diminishes the ability of the pendulum to 
resist the other disturbances, in an escapement (47) in 
which these other disturbances are much reduced, while the 
dead friction is much increased. Indeed it is obvious 
without either calculation or experiment, that the more 
friction there is, the larger must be the impulse required 
to make up for it, and consequently the larger wiU be all 
the errors connected with the impulse. 

37. We see now the true cause of the accuracy of the 
dead escapement, and also how we are to set about it, to 
make this accuracy as great as possible. For though we 
camiot determine the proportion which the increase of the 
arc bears to that of the force, since it depends upon the 
varying friction of the different parts of the clock, yet we 
see that they have a tendency always to correct each other ; 
and whenever the state of the clock is such (as in one 
experiment I actually found it to be) that the arc increases 
just one-third as much as the clock-weight is increased, 
those two parts of the error will exactly counteract each 
other : and the ratio of the increase might happen to be 
such as to compensate the circular error also. It generally 
happens, however, that as the clock gets dirty, the force 
and the arc decrease in such a proportion that the loss of 
time preponderates. 

But there is one case in which the opposite effect 
not unfrequently takes place to the surprise of those who 



know the common result of a decrease of arc, and of 
which for some time I could not myself discover the rea- 
son. Church clocks will often be found in a few months 
after they are put up to increase their arc of vibration con- 
siderably, and at the same time to ffain. This increase of 
arc arises chiefly from the decrease of friction on the dead 
part of the pallets, owing to the teeth and pallets polishing 
themselves more perfectly than had been done by the 
maker. Moreover in most clocks the quantity that we 
have called D is much larger than in the above example ; 
and therefore the terms depending on D are much larger 
than the term containing the circular error. Consequently 

the term D may preponderate over both the. other 

terms ; and as it has a — sign the clock wiU then gain 
while the arc increases. 

It may be as well to explain that when a clock gains it 
is said to have a + daily rate of so many seconds, and. 
when it loses, a — rate; and it should be remembered that 
these §igns are the reverse of those which indicate the 
decrease or increase of the time of an oscillation. I may also 
remind the inexperienced reader that the goodness of a 
clock is indicated, not by its rate, but by the variation in 
its rate. 

38. The effect which I have just now mentioned, pf the. 
self-polishing of the pallets is of course only temporary ; 
and the general effect of a decrease of force and of arc in a 
4ead escapement is, that the clock gains a little, whereas a 
common recoij escapement loses considerably as the arc 
decreases; and this has led,, to the ado^ion of a smaU 
recoil in the place of the dead part of the pallets, especially 



in turret clocks^ whicli are more liable to changes of 
friction than others. This recoil may be given hj striking 
the circle of the dead part of the pallets, not from the axis 
of the pallets, but from a point a little below that axis, in 
the line of centres of the pallets and the scape-whed, which 
produces a circle with a higher degree of curvature, and 
therefore raises the teeth a little after they have dropped 
onto the pallets: and the farther the pendulum swings 
the greater is the degree of recoil. This is commonly 
called .the half^dead escapement. 

89. There are some other things to be learned from the 
above expression for the daily error of a dead escapement, 
and some of them apply to all escapements. First, we 
observe that, as the weight and length of the pendulum are 
ia the denominator of the. fraction, the larger they are the 
less the errors of the clock will be; and this is the case 
with all escapements, and with all the errors, Whether of 
friction or anything else, connected with the escapement; 
for it is quite a vulgai j^rror to suppose (other things being 
equal) that a heavy or long pendulum requires a heavier 
dock weight than a short and light one, except that of 
course in large clocks the pendulum spring is stiffer, and 
the whole of the machinery heavier, and so requires a 
larger weight to move it than in small ones. The best 
turret clocks have 2, or 1^ seconds pendulums, which are 
about thirteen and eight feet long respectively, and the bob 
sometimes as much as four cwt. Now if the same clock 
had a one second pendulum with a bob of half a cwt., 
instead of a two seconds pendulum with a bob of four cwt., 
it would go just thirty-two times worse ; or in other words,,. 


if with the long and heavy pendulum it varied a minute in 
a months with the other pendulum it would vary above half 
an hour. 

40. The next independent quantity in the equation far 
B or dJ) is (/— c, or the diJBferenee between the angle at 
which the impulse begins and the angle (on the other side 
of zero) at which it ends^ which is called the af^le of 
escape. And we see that the smaller this difference can be 
made the smaller all the errors of the escapement will b^ 
as indeed was apparent from the general reasoning inde- 
pendently of the exact value of D. The limit to the 
smallness of d-^c is merely a practical one, depending upon 
the accuracy of the construction of the escapement, and 
upon the length of the pallets (by which I mean the arms 
that carry the pallets), for of course the longer they are the 
smaller will be the angle c^—c corresponding to a given 
linear space required for the tooth to fall upon. The 
length of the pallets can only be varied in two ways ; either 
by increasing the wheel and pallets together, or by in- 
creasing the number of teeth which the pallets embrace. 

41. But it is to be observed that as you increase the 
length of the pallets you also increase the linear space over 
which all the friction acts, both the friction during the 
impulse and the dead friction ; and if the wheel remains 
of the same size, the pressure, to which the intensity of 
ihe friction bears a certain proportion, will be the same ; 
aad so the total quantity of the friction will be increased 
by increasing the number of teeth embraced by the pallets 
in order to increase their length ; for the radius of each 
pallet, drawn from the pallet arbor to the comer of the slope. 


must be a tangent to the circumference of the teeth of the 
i^cape wheels in order that the teeth on each side may act 
equally^ and for the same arc of the pendulum^ upon the 
slopes of both the pallets. The pallets therefore are seldom 
made to embrace more than the space of 10| teeth^ or 11^ 
at the most^ when the wheel has thirty as usual. 

4t2. Again it is not desirable to increase the size of the 
wheel beyond what is necessary for its proper strength in 
proportion to the number of teeth, because its moment of 
inertia increases even more rapidly than its size, and so 
causes it not to follow the pallets so quickly : and when 
tithe wheel is too large, the teeth may often be heard to 
jump or chatter on the pallets, from the length of the drop 
land the great Hnear velocity they have acquired when they 
are suddenly stopped. In astronomical clocks, or 'regulators' 
as they are called, the scape-wheel is generally two inches 
in diameter or a little more, and cf^-c can be made as little 
as £0' and ought never to be more than 30'. In turret clocks 
five inches is quite enough for the diameter of a scape 
wheel with thirty teeth ; and as that will allow more than 
twice as much linear space to the same angle as a wheel 
only two inches in diameter, c'— c, may be made as little as 
in a regulator. I have improved the going of a church 
clock, which was very weU made in other respects, by opening 
the pallets wider, as they had been so set that the tooth 
fell on the circular part a good way above the slope, in- 
stead of as near to it as possible; and they fortunately 
admitted ef adjustment. 

43. I have said nothing about the size of tZ-f c as well 
as d—c, because that quantity does not enter into the 


value of D. But it is evident that the larger these angles 
are, the longer the impulse will last and the less dead 
friction there will be for the same degree of oscillation; and 
also the less suddenly the teeth have to drop, and conse- 
quently the more closely they will follow the pallets at the 
beginning of the impulse. But on the other hand the 
friction occurs during the impulse instead of occurring on 
the, dead part; and the question between long and short 
impulses is one which is perhaps better determined by 
eKperiment than theory; the result however both of theory 
and experience seems to be in favour of a short impulse^ 
especially as it requires a less arc upon the whole, and less 
maintaining force. Mr. Dent makes d^ the larger of the 
two angles not much above half a degree, and the angle of 
oscillation 1|° in his astronomical clocks, and some of them 
are going with only a weight of four lbs., and the pulley, 
faUing three feet in the eight days. A clock of this sort 
is also . Bafer than one in which the angle of impulse is 
nearly equal to the whole arc of vibration, for in that case 
a little diminution of the arc &om any accidental cause, 
such as freezing of the oil, will cause the clock to stop. 
This was what the old dockmakers meant when they said 
that the excttrsion, or the excess of the anglp of vibration 
above the angle of escape, ought to be large. 

44. There remains to be noticed one other ingredient in 

the equation for D, viz. — r; for these quantities evidently 

depend on each other, as you cannot increase the arc in a 
given clock, but by increasing either W the clock-weight, or 
A its fall in the day. As was observed before, it is impos- 
sible to say practically what increase of arc ^ a any given 


increase of WA will produce ; but it is certain that the arc 
increases much more slowly than the weight ; and^ moreover^ 
as the arc is increased, the quantity of dead friction is in- 
creased; and, therefore, on the whole, it is found that no 
good is done by adding to the weight in order to increase 
the arc. (This remark will not apply to the next class of 
escapements in which there is no dead friction.) But if 
you can increase the arc by diminishing the friction on the 
pallets, or improving the suspension of the pendulum, that 
is a clear gain; and accordingly the less weight and fall a 
dock requires to make the pendulum vibrate a given arc 
(other things being equal), the better the clock will go. 
For this purpose the pallets ought to be made as hard 
as possible ; and in highly finished clocks they are made of 
j.ewels ; sapphires I understand are the best. It seems to 
be a question whether steel or brass teeth work best upon 
jewels : upon steel pallets there seems to be no doubt that 
brass teeth work with less friction than steel and require less 
oil ; but the brass should be hammered so as to make it as 
hard as possible. 

45. Various contrivances have been proposed for di- 
minishing the amount of dead friction, such as having 
separate pallets for the dead part of the action, which are to 
be left behind by the pendulum as it advances beyond the 
angle of escape (c/), and carries the impulse pallets with it. 
It will be seen, when we come to renwntoire escapements, 
why this plan is objectionable, besides the difficulty of con* 
structing it. Perhaps it might answer to put a large and a 
small scape- wheel on the same arbor, and short and long 
pallets on the same pallet arbor, the small wheel to give the 



impnlse on the long pallets, aod the teeth of the lai^ whed 
to be stopped by the 
Bhort palleU contain- 
ing only the dead 
part, which however 
is to have a small re- 
coil in it. Inthatcase 
the impulse might be 
given throngh a small 
' angle n&a the middle 
of the vibration ; and 
the presaate which 
causes the dead fric- 
tion wonl^ be less 
from a lai^e wheel than a small one, and the space to be 
travelled over on the dead part of the short pallets wonid 
also be less than on long pallets; so that iu every way 
there would be an advantage as r^ards Miction. This may 
be called a di^Ux escapement, as it agrees with the escape- 
ment of that name in watches. It would lequire con- 
siderable accuracy in its construction, bat not more than 
mother, which I shall mention presently, as having been 
constructed, and at work.* 

46. There is mother form of the dead escapement, 
which does not differ in principle from the common one, 
but has some mechanical advantages over it, espedally for 
U^e clo<^. It is called the j»»-w^e/ etcapement. It will 

• I hare made thia and several other drawinga in this book mote 
with a view to an intelligible exhibition of the action of the puts, 
than to their aetiuJ at proposed oonatrnction. 


be snfQciently clear from the drawing, that the pins are set 
on the face of the 
scape wheel in- 
stead of teeth on 
its edge, and that 
the two pallets, 
instead of embra- 
cing about one- 
third of the cir- 
cnmference of the 
wheel, are pat so 
near together as to 
leave roomforonlj 
one pin to pass 
between them ; and 
the end of one 

slope shonld be just over the beginning of tiie other. 
The pins are only semi-cylinders, since the npper part of 
the cylinder could not act, and cntting it away allows 1^ 
pallets to slip throngh close above the teeth, so as to waste 
as little drop as possible. The advantages of this escape- 
ment are, first, that it does not require so much accuracy 
of construction as the oth^; for in the common one if every 
tooth is not exactly in its right place, with reference to eveiy 
other at the distance of ten or eleven &om it, the escape- 
ment is liable to stick, and if the clock is going with a 
heavy pendulum a tooth is then pretty sure to be broken ; 
whereas if every successive pin in the pin-wheel escapement 
is neaily at the same distance from the one immediately 
before it they are sure to cleat the pallets : secondly, if by 


any accident a tooth of the common scape wheel is broken 
the wheel is ruined ; whereas if a pin is broken a new one 
can be put in in a few minutes : thirdly, many more pins 
can be put into a wheel of given size, so as to clear the 
pallets, than teeth of the usual shape ; and therefore there 
is less drop and waste of power at every beat,* and the 
wheel turns through a less angle and with less velocity, 
and therefore with less friction on its pivots, and can also 
have a larger number of leaves to its pinion^ the advantage 
of which will be seen when we come to consider the wheel 
work : fourthly, both strokes on the pallets being down- 
wards, instead of one downwards and the other upwards, 
there is less shake in the pallet arbor; in a common escape- 
ment the difference of stroke on the up and the down 
pallet can generally be distinguished by the sound. 

The only disadvantage of this escapement, as for as I 
know, is that the force of the cylindrical pins on the palleta 
is not so uniform from the begmning to the end of the 
impulse as with sharp teeth and pallets with straight slopes;- 
in fact, the slopes ought strictly to be concave, in order to 
make the inclination of the tooth to the pallet the same at 
the end as at the beginning of the impulse. This however 
oould not be done without the introduction of greater evils 
than the very small variation in the force. The pin escape- 
ment has been long used by the best makers of turret 
docks both here and in Erance. It is not used in 
the Exchange dock, because it is not so well adapted for 
jewelled pallets, which that clock has. Mr. Yulliamy 
makes the pins of steel acting on broad pallets, portions of 
turned cylinders, and without any recoil ; diminishing the 



radius of the cylinder^ or setting the same pallets on longer 
arms^ would give them a small recoil. Mr. Dent uses hard 
bra^s wire pins acting on pallets, not flat, but having the 
cross section a segment of a circle, and he makes the escape- 
ment ' half dead.' The scape-wheel for a 1 ^ sec. pendulum, 
with forty pins in it, in the two clocks I shall mention, is not 
quite four inches in diameter ; this shows the small amount 
of drop which this escapement requires compared with the 
c(»nmon one; 


47. Before I proceed to the next class of escapements, 
and by way of introduction to them^ I will describe the one 
I alluded to just now, which was invented by the astronomer 
royal, and of which three or four specimens have been 
made by Mr. Dent : one of them is now going in his 
shop in Cockspur-street. In order to prevent the inequali- 
ties of force of the train affecting the impulse on the pendu-. 
lum, there are two scape-wheels and two pairs of pallets, 
one for the stop, and the other for the impulse : the stop- 
wheel is the one connected with the train, and the impulse- 
wheel rides on the same arbor, and is connected with the 
other by a spiral spring. The stop-wheel is let go by its 
pallets, which have no sloped £aces, just before a tooth of 
tiie impulse-wheel would arrive at the slope of its pallets^ 
and so the tooth is carried down the slope, and the impulse, 
given by the force of the spring only. In fact, if the reader 
turns to the drawing of the duplex escapement in page 64, 
and supposes those two wheels reduced to the same size,. and. 


connected by a spiral spring instead of by screws, it will re- 
present this escapement (which I may observe is not the one 
suggested by Mr. Airy some years ago in his paper in the 
'Cambridge Philosophical Transactions^). The advantage 
of it is that the impulse is constant, or at least has no 
greater variation than that of the force of the spring arising 
from changes of temperature, which is much smaller than 
the variations in force caused by the friction of the train ; 
with however this not immaterial exception, that the 
impulse-wheel tiims with more friction, riding on the arbor 
of the other wheel, than if it turned on pivots as usual. 
Moreover the dead friction is that due to the train, and 
is very much greater than usual. The reader may form 
sameTome idea of the amount of force consumed by this 
friction and the additional weight of the second scape- 
wheel, when I state that (reducing the falls to the same 
amount) the weight employed in driving this clock is more 
than twice as much as that employed in driving one of Mr. 
Denf s second-rate regulators, vibrating the same arc, and 
more than three times the weight of one of his first-rate 
regulators, also with the same arc; notwithstanding the 
numbers of the pinions of the clock with the duplex escape- 
ment are higher than those of either of the others, and the 
dock is made with the utmost care. And it is not sur- 
prising that this large amount of friction more than coun- 
terbalances the advantages of this escapement, and the 
dock does not go so well as a first rate regulator with the 
common dead escapement. Both these objections might 
probably be diminished; the first by making the stop whed 
ride upon an arbor or stud set on the frame, and carrying the 


pivot of the impulae wheel in the maimer I shall have to 
describe in a chnich clock lately made according to some 
sn^estions of my own (177) ; and the second by n9ing a 
large and light stop-wheel with short pallets in the manner 
above described (45). 


48. Escapements of the kind jnst now described^ (so 
far as the impulse wheel is concerned) in which the impulse 
is given to the pendulum by a small separate weight or 
spring, independently of the force of the train, are called 
by the Erench remantoke escapements, because the clock 
train winds or lifts up the maintaining force at every beat 
or at some given number of beats of the pendulum. They 
have long exercised the ingenuity of clockmakeis; so 
long indeed and generally so unsuccessfully^ that they 
appear to be considered by some people the philosopher's 
stone, or the perpetual motion, of dockmaking. It would 
be impossible to describe in any reasonable compass the 
various inventions that have been made for the purpose, 
both for clocks and chronometers. Eor chronometers they 
have always hitherto faUed, and I have no doubt they 
always will; not only on account of the excesssive trouble 
and difficulty of constructing them on so small a scale, but 
because a chronometer train is so ]ight that there is npthing 
like the same friction and waste, and therefore variation, of 
force between the main-spring and the scape wheel as there 
is in clocks, especially turret clocks, for which remontoires 
are most required. But that they can be made and ill 
answer for clocks, both large and small, is fully proved by 


Mtmul examples of them which I shall describe as actually 
now at work with complete snccess, hesides the one which 
I JDBt DOW described and which may be considered an 
ewiq)ement of the ' transitioii style' between dead and re* 
montoire escapements, or rather between an escapement 
remontoire and a train remontoire. 

49. When the impulse is given to the pendulum by a 
■mall weight or weights, instead of a spring, raised at every 


beat or other int^rval^ thej are called gravity remontoires^ o)r 
merely gravity eacwpemmU. 

The most simple^ though not the earliest form of the 
gravity escapement is this. A G^ B C^ are two arms turning 
separately, on pivots at C which coincide as nearly as 
possible with the axis of suspension of the pendulum: 
At the lower ends of the arms are two pallets of the shape 
in the drawings so that a tooth of the scape wheel will 
slide along the sloped part and raise the arm until it comes 
to the little detent or hook at the end^ which stops the 
tooth : the tooth at B is here represented as stopped^ and 
the tooth at A reieuly to raise the arm A€^ as soon as the 
other arm. is pushed farther out so as to set the wheel free. 
The arms liave projecting pieces, D, E, reaching down to 
the proper distance to be met by the pendidum rod. The 
pendulum is here drawn as going to the right and just 
touching the projecting piece E of the arm B C : as the 
pendulum goes on it will raise that arm and set the 
wheel free to raise the other arm; and the pendulum 
will carry the arm BC with it as far as it swings, and 
when it descends again the arm will descend with it, 
not only as far as the place where it was taken up, but 
farther, that is, until the slope of the pallet B rests upon 
the next tooth of the wheel, or upon some fixed stop set in 
tiie frame at the proper height. The maintaining force 
upon the pendulum depends therefore, first, upon the 
weight of the arms, and secondly, upon the difference 
between the angle of the pendulum when it takes up each 
arm in ascending and leaves it in returning. If the arms,- 
instead of. acting by their own weight, were so counter-^ 


balanced that their weight did not act upon the pendnlnm^ 
and were fixed b j short and rather stiff springs like a 
pendulum spring at C, the action would be evidently just 
the same, only it would then be a ^prinff, instead of a 
gravity remontoire; and if not counterbalanced, it would 
be a compound of the two. 

50. This simple form of the escapement will do as well 
as any other to ei^lain the general advantages, and the 
mathematical conditions to be observed in the construction 
of all these escapements, whatever may be their mechanical 
peculiarities. The advantages of them are evidently these : 
first, the impulse depends upon the action of a given 
weight pressing on the pendulum through a given distance 
(that is a given difference between two distances) and 
communicated without any friction, except the incon- 
siderable friction of the pivots at G ; the force is therefore 
independent of all variations in the friction in the train 
and escapement. Secondly, there is nothing corresponding 
to the dead friction of the dead escapement, or the still 
greater friction of the recoil escapement; and therefore the 
pendulum will be much less liable to variations in its are 
of vibration : indeed there is no friction at all except the 
momentary friction of unlocking the teeth when the pen- 
dulum first catches the arms ; and therefore the pendulum 
will swing a given arc with less nudntaining force. A 
third advantage has been supposed to be that the pendulum 
may be leffc quite free for some distance during the middle of 
its arc, since it need not be nutde to take up one arm as soon 
as it leaves the other : this however we shall find to be the 
worst construction, though it has been the most common. 


51. Now if we could be sure that the pendulum would 
never vary at all in its arc, this escapement would be 
mathematicallj perfect. But there is no such thing in 
practical mechanics as perfectly invariable motion ; and it 
unfortunately happens that if the arc of vibration varies at 
all, from change in the density of the air or the little 
friction which exists, it produces much worse effects with 
this escapement than a much larger variation of the arc 
produces with the dead escapement. Por as was shown in 
the recoil escapement, the additional force, or weight of the 
arms, acts with gravity both in the ascent and descent of 
the pendulum, and therefore the farther it ascends with 
this additional force acting upon it, the more the time of 
its oscillation will be diminished. In order to show the 
amount of this acceleration, let D as before be the' increase 
of the time of aU the oscillations of a seconds pendulum 
in a day when attached to this escapement : c the angle 
(from zero) at whict the pendulum takes up each arm, c^ 
the angle at which it leaves the other arm ; only you must 
remember that d is now supposed to be on the same side 
of zero as c, and not on the opposite side as in the dead 
escapement ; "W, h, M, l, and a, indicate the same things as 
in the dead escapement. Then it may be proved that 

The — sign indicates that D is here a decrease of time 
instead of an increase, c and d may be any angles we please, 
except that d must be less than c, or there will be no 
impulse given to the pendulum, and also c cannot safely be 
more than a — 30', since the unlocking of the wheel has 


to take place while the pendulum is moving through a^-^e. 
It will be found that whatever value of c— (?* we take, 
subject to these conditions, the value of D in this escape- 
ment is many times greater than in the dead escapement. 
And since D is so much larger in the remontoire than in 
the dead escapement, it was perhaps natural to suppose 
that the variations of D (which are the error of the clock) 
must be larger abo; akd experience was supposed to confirm 
this theory, for however carefully the remontoire escape- 
ments have been made, they have not generally equalled 
the accuracy of the dead escapement, and when they have, 
only by their mechanical advantages and veiy perfect con- 

62. But although this conclusion would be perfectly 
correct if applied to any construction of the remontoire 
escapement which a clockmaker would spontaneously adopt 
(which accoTmts fot experience appearing to confimi theory), 
there is nevertheless a construction, wliich, though not 
very convenient^ is quite practic^-ble, and will render the 
variations of D so small as to be inappreciable for any 
probable variation of the arc of vibration. In order to 
produce this effect, it may be proved that <?, d^ and a must 
be made to satisfy this condition : 


Many different values of c and d for any given value of a 
will theoretically satisfy the equa,tion ; but practically the 
number-is very limited by the circumstances mentioned a 
little while ago. For suppose that we intend a to be 12(K 
and ff to be on the same side as c; then in order that a — e 

bemontoibj: escapement. 75 

may =£ 80'^ c must not be more tha^ 90'^ and we shall find 
that the corresponding value of ^ is very nearly 78'. But this 
construction is barely practicable^ for the angle through which 
the impulse has to be given^ and through which the arms have 
to be raised by the escapement will be only 12', and these 
small angles are extromdy difficult to manage with a<^uracy 
in the construction ;9f the escapement. Probably the only 
way in which it can be done is by such a construction as 
that which many persons must have seen in the window of 
Mr. Denf s shop, made^ however, without any reference to 
this theory, in which the action takes place near the bottom 
of the pendulum, and so the linear space corresponding 
to a small angle is sufficiently large. This construction, 
however, is on another account very inferior to that in 
.which e and </ are on opposite sides, or the arms act upon 
the pendulum through a considerable angle on each side of 
sero; as may be seen at once from the fact that, in that case, 
e — (f, which occurs in the denominator of the value of D and 
its variations, becomes e+(f, and therefore the value of 
D and dD is very much smaller than when c and cf are 
on the same side of zero, or the pendulum is free in the 
middle of its swing. 

SS. Suppose then that e and {/ are on opposite sides, and 
a=120' as before; then one construction that will answer 

is to make c and & each =-^ or 84'* 86, so that one arm 

is always taken up just as the other is left. And as it 
is not safe to allow the pendulum to touch one arm before 
ihe other has reached its lowest point, and is ready to catch 
the scape-wheel, any value of </ larger than c is practically 
inadmissible ; and on the whple the best form of the remou- 


toire escapement evidently is that in which c and ^ are on 
opposite sides of zero, and each = y^ or *707a. In that 
case the equation in § 51 assumes the more simple form, 

and the variation of D toi das, given small variation of a, 

,T. Wh a* da 

M/iTfl* /^ , a 

Now in order to see what the ^ror of such a dock will 
amount to for certain small variations of the arc, we may 
put for W^ half its value in the deadeseapement, as that is 
sufficient for a weU-made clock of this sort, there being no 
dead friction to overcome ; and as a = 120' and therefore 

c = 84'-86, or -0246, m71^ = ^'^ > ^^ ^® shall find 
that if a increases or decreases from 120' to the followiog 
amounts, the clock will in either case gain, but no more than 
the following small quantities daily : 

a = 123' dD = Yh ^^ ^ second 

122' ^ 


118' ^ 

117' ^ 

Therefore as the arc will never spontaneously increase, 
it should be so adjusted when the pivots and pallets are 
cleaned, that a may exceed the normal value by a little more 
than 1', which it will do if the pendulum is made to swing 
to 120' in a clock in which c is 84'; and then the arc 
may diminish as much as 4/ or 5' without producing any 


aensiblQ effect upon the rate of the dock ; and sach a clock 
may be pronounced theoretically perfect. 

If c should be made a more convenient size^ say 1^ for 
a = Z^, the clock will lose for an increase and gain for 
a decrease of arc : but it will only lo^ 4- of a second a day 
for a decr^ oi 5', wHch is a very large variation of 
the arc for such a dock ; and^ therefore^ even with this con- 
struction^ it would probably go better than any dead escape- 
ment. These escapements have however, I believe, been 
always made on the plan of leaving the pendulum free 
in the middle of the swing, which, as we have seen, is 
the worst construction, especially as the angles were sure to. 
be made veiy far wrong, from the difficulty of making them 
right on that construction. 

54. There is however a certain mechanical difficulty in 
the construction of these escapements, which has probably 
been a greater obstacle to their use than the supposed 
mathematical objection. It will be easily perceived on 
looking back to the drawing of the escapement that there 
is some risk of the teeth of the scape wheel, when driving 
up the slope of the pallet, sending the arm too far by 
causing it to rise too quickly ; and if it does, the hook at 
the end of the slope will not catch the tooth as it ought to 
do, and two or three teeth will slip past at once : this is 
called tripping. Various contrivances have been resorted 
to to prevent it; the most obvious is to put the slope on 
one arm, and the hook on another by the side of it j the 
arm with the hook is not allowed to fall so low as to require 
raising by the wheel, and so is always ready to receive a 
tooth when it is not raised by the pendulum; the pendulum 

78 cumming's escapement. 

raises both the hook-arm and the pallet-ann together. This 
is Cunmung's escapement, with the omission of some halls 
which he added by way of giving ' momentum' to the arpis, 
in complete ignorance of mathematical principles^ that being 
the very thing we ought to avoid as much as possible ; and 
with the arms set upon springs instead of turning On 
pivots and carrying weights, it is Hardy's escapement^ 
which the transit clock at the Cambridge Observatory has. 
But this is a compUcated arrangement^^nd I believTevea 
these stationary hooks (as they may be called in contrast to 
the others) sometimes trip from tiie blow of the tooth 
against them, unless they are nndercut, or the teeth so 
made to fit the hooks that they cannot be disengaged by the 
pendulum without causing a slight recoil in the wheel and 
a resistance to the pendulum, which is of course very ob- 
jectionable, especially as the force required for it will vary 
m different states of the clock. It wiU now be seen why 
the similar contrivance which I mentioned (45) as having 
been appUed to the dead escapement does not answer; for 
the separate arms carrying the dead part of the pallets are 
liable to these same objections, and moreover on account of 
their own weight they affect the pendulum after the manner 
of a remontoire escapement without the angles properiy 


55. A form of the gravity escapement has lately been 
invented by Mr. Bloxam, the inventor of the dipleidoscope 
(9), which when it is so made as to satisfy the proper con- 
ditions respecting the angles, (which he had unfortunately 


bot discovered until after he had constructed the escape- 
ment withoat an; particular regard to tiie angles), seems 
likely to possess eret; qnaMcation that is requisite to 
produce s perfect escapement, and with no inordinate 
difficulty of construction. By Mr. Bloxam's pramission 

I ^Te a drawing, and a short description of it, as it 
wiU require to be made. A.C, BC are the two arms. 



carrying flat stops corresponding to the common hooka 
at A, B, one of which is represented as stopping a tooth 
of the scape wheel at B. I have omitted the pendulum 
in the drawing for greater distinctness. The scape wheel 
has only nine teel^^ and consequently moves through 
20° at each beat, and its motion therefore will take a Ion- 
ger time, or will be more gradual, than if it only moved 
through &° as usual. Concentric with the scape wheel and 
fixed to it are two other wheels of the form called cam 
wheels in machinery, though no great accuracy is required 
in the form of the teeth for this purpose. These wheels are 
so placed that one of the nine teeth or cams of the smaller 
wheel is ready to raise the arm B C by means of the pro- 
jecting piece E when that arm is left by the pendulum at 
its lowest position ; and while the wheel turns through 20** 
the small wheel is to raise the arm B C through Si', if the 
pendulum swings 2°; and in like manner the larger cam- 
wheel raises the other arm through the same angle, by 
means of the projecting piece A B. 

The reason why two cam-wheels are required, in order 
to completely satisfy the condition respecting the angles, 
is, that as one arm must be raised by the lower part of the 
wheel and the other by the upper, they cannot both be 
carried through the same angle, unless the wheel which has 
to work the longer arm is larger than the other. 

Now the effect of this arrangement is, first, that the 
arms are raised with much less friction than when a tooth 
slides alonjsc a sloping pallet: secondly, the rise taking a 
longer time! there is no risk of the a^s being driven too 
far by their own momentum; thirdly^ the pressure of the 

MR. bloxam's geavitjt escapement. 81 

teeth of the scape-wheel upon the stops is much less than 
in a wheel of the uaual number of teethj^ and consequently 
the Mction at unlocking is mucKless ; in fact the pressure 
is so little that tho stops^ inst^ of being undercut^ may 
be sloped a little the other way without the pressure being 
sufficient to lift them^ and therefore the £riction at unlocking 
may be reduced to nothing. To show how safe from trip- 
ping this escapement is, I have seen the weight in Mr. 
Bloxam^s clock pulled by hand, and so increased to more 
than double of what was required to work the escapement, 
without its exhibiting the least tendency to trip. And it 
may be observed, that even if the angles should not be 
quite correctly adjusted, the causes of error are so much 
reduced by the absence of nearly aU friction, that hardly 
any variation of the arc can take place: in fact Mr. Bloxam 
tells me that he has never been able to detect a variation in 
the rate of more than a second a week even in his clock as 
now made with the angles of escape not half the proper size. 
1 have no doubt it might easily he made on a large scale 
with perfect accuracy and no great amount of trouble, and 
it might also be combined with a remontoire in the train, 
for the purpose of effecting those objects for which a train 
remontoire, letting off about every half-minute, is desirable 
in tunet clocks, independently of their accuracy of per- 
formance, for which see § 172, &c. It may be convenient 
to state the proportionate dimensions which I find wiU be 
required for the different parts. Suppose we intend the 
radius of the scape-wheel, (which is very light, being cut 
out of a plate of thin steel) to be 2'5 inches, then we 
must make the other parts as follows :— 



Eadius of larger cam-wheel 1"24\ 

Badius of smaller cam-wheel 1* J supposing the 

Distance from centre of wheel > angle a = 120 

to axis of arms ... 7*45 \ and c = 84' 

Length of arms (to pallets) 7* / 
These dimensions will allow -^ inch for the depth of loddng- 
if it occupies an angle of 24'^ and iV if i^ occupies as much 
as 30': anything between the two will do very well. If the 
scape-wheel is half this size everything else must be half 
the size. The counterpoises a, 6, are added because I 
believe the arms cannot be made light enough to do with-^ 
out, when they act through an angle of 168'. 

56. Other escapements without number have been 
invented: indeed there is a story of a celebrated clock- 
maker saying he would undertake for a wager to invent 
a new one every day. But a description of them would be 
of no use in a merely practical treatise, as the escapements 
themselves have never come into use. Perhaps the most 
ingenious, as well as curious, was Harrison^s, who, when he 
was only a carpenter, invented it to save himself the trouble 
of having to go so frequently to oil the escapement of a 
turret clock, which he had undertaken the care of. It has 
no friction on the pallets, but has an immense recoil; and 
though a degree of accuracy is attributed to one made by 
him, which is evidently fabulous, it is said (which is pro- 
bably not fabulous) that nobody else could ever make them 
to answer. A description of it may be seen in several of 
the Encyclopaedias, and in Beid's ' Treatise on Clock- 
making.' We will proceed to consider another matter, which 
though appareaily piinute is of great importance. 



57. All the substances of which a pendxdum rod can 
be made increase in length as their temperature increases. 
Let I be the length of the rod, d I its increase for any given 
number of degrees of heat, and d t the corresponding increase 
of t, the time of the vibration of the pendulum, then (re- 
membering that dl \& Y^rj small compared with /, and so 
f Y) iJ^y l>e neglected), 

—t 77 ^-^u' ^'"^^ Tl- 2 

if for shortness we put m for j the rate of expansion of the 

material of the rod for some given number of degrees of 
heat. And the dailj loss of the clock, which we may call 
d T, will be 43,200 m in seconds, whatever is the length 
of the pendulum.* The following is a table of the value 
of i», or the rate of expansion in length of the following 
materiab, for 10° of heat : — 

Wyte-deal - - - *000023 

Hint glass - - - • 48 

Steel rod - - - - • .... 64 
Cast iron ... '....GG 

Lronrod - - - - '...•7 
Brass - - - * • ... 10 

Lead -...16 

Zinc • . - - •...17 

Mercury (in buUc» not length) ' . . 100 

* This same calculation will show us how mtich a pendulum ought 
to be shortened when it loses, or lengthened when it gains — assuming 
the weight of the rod to be immaterial compared with the bob. Sup- 
pose for instance, that 33 threads of the screw are contained in an 
i&oh, and that the whole length / of the rod is 45 inches ; then each 


Thus for a common pendulum rod of iron wire, we 
see that dT for an increase of temperature of only 
10^=43200 X -00007=3 sec: or the clock will lose a minute 
in three weeks ; and if the pendulum is adjusted to go right 
in winter, it will lose about a minute a week in summer. Even 
a deal pendulum would vary nearly a third as much as this.^ 

58. We want therefore some contrivance which wiU 
compensate this expansion of the rod; that is, which wiU 
always raise the centre of oscillation of the pendulum as 
much as the expansion of the rod lets it down. If the rod 
had no weight, and the bob were merely a heavy point, this 
would be the same thing as saying that the centre of gravity 
of the bob must be kept at the same height ; and as the bob 
of compensated pendulums always is heavy in proportion to 
the rod, this is approximately true, and with the addition of 
a small quantity, according to a simple rule, it is sufficiently 
correct for practice. 

59. The most simple kind of compensated pendulum is 
one in which the material of the bob expands so much more 

thread, or each complete turn of the nnt, ivill raise the bob by a 
quantity = -rjjij-, which we may, as above, call m; and the corres- 
ponding daily alteration in the time being = 43,300 m, will be just 
30 seconds. Consequently, if the head of the screw is divided, as it 
usually is in regulators, into 60 divisions, with a pointer over them 
a turn of the screw one division to the right or left will accelerate or 
retard the clock half a second a day. 

* You must not expect to find this result actually take place in a 
common house clock : the other causes of disturbance in such clocks 
are so large, that they may either overbalance or aggravate the effects 
of heat upon the pendulum rod. I believe they will generally be of 
the former kind, since the heat makes the oil more fluid, which 
in the common recoil escapement will accelerate the clock. 


than that of the rod^ that a bob of moderate length resting 
on the bottom of the rod will raise its own centre of gravity 
as much as the expansion of the rod lets it down. A deal 
rod^ with a leaden bob about f ths of its lengthy will be thus^ 
compensated. For the ratio of the expansion of the wood 
to that of the lead is we see about -f ; and consequently 
a bob whose centre is |th of the length of the rod from its 
bottom, will compensate the rod. 

60. But we want to know how much longer than the length 
of the simple pendidum the rod must be in order to carry a 
long enough bob. Let l^ as before, be the simple pendulum, 
or (as we assume it to be) the distance from the point of 
suspension to the centre of gravity of the bob ; x the addi- 
tional length required, or half the length of the bob ; m the 
rate of expansion of the rod, n that of the bob. Then 
(/+«) m^ the expansion of the rod downwards, must=;i? », 
the expansion of half the bob upwards; ora? = ^7-. This 
calculation will make the leaden bob of a 39 inch deal pen- 
dulum about 13 inches long, and the rod 45^. It is foimd, 
however, on account of the difference between the centre 
of gravity and the centre of oscillation, that the proper 
length is 14^ inches; and the practical rule may be given, 
to add ^th to the length of compensating material deter^ 
mined by calculation on the hypothesis of the centres of 
gravity and oscillation being identical. 

61. The principle of the mercurial pendulum is exactly 
the same as that of the wood and lead. The rod is of steel, 
and the mercury is put in a cast-iron cylinder (in the best 
pendulums) screwed to the bottom of the rod. Only it is 
to be remembered that the rise of the mercury in the 


cylinder will be diminished by the lateral expansion of the 
cylinder itself^ and consequently a rather greater height 
of mercury is required than that given by merely taking 
the tabular rate of its expansion. The old form of mercu- 
rial pendulum was that of a glass cylinder standing on 
a itirrup at the bottom of the rod. The chief advantage of 
the iron cylinder is that it can be made of a more regular 
shape^ and that it takes the same temperature as the rod^ 
and communicates it to the mercury more rapidly than the 
glass. Captain Kater^ in his chapter on compensated pen- 
dulums in 'Lardner's Mechanics' (from which the above 
table is taken)^ says he has successfully employed, as a 
cheap mercurial pendulum, one made entirely of glass, the 
rod and cylinder being blown in one piece. The height of 
mercury required in an iron cylinder is stated to be 6'6 
inches. The best mercurial pendulums are actually tried 
ftnd adjusted for compensation at various temperatures, by 
adding or taking away mercury as may be required. 

62. But mercurial pendulums are too expensive for 
common use, and would cost nearly as much as the dock 
itself, for pendulums weighing several cwt.; and wood 
is open to the objection that it is liable to twist, and can 
never be completely protected from damp, which of course 
alters its weight ; and therefore some other kind of metallic 
compensation is necessary. Now it will be seen on looking 
at the table, that the ratio of the expansion of steel to that 
of brass is *61. Consequently, if we can make a pendulum 
of steei and brass rods in the proportion of 1 inch of steel 
to * 61 of brass, and so arranged that the brass rods carry 
the bob up while the steel ones let it down, it will be com- 


pensated. The only convenient way of doing this, is to 
make the brass rods (for there must evidently be a pair 
of them^ one on each side of the main rod) rest on a 
nut at the bottom of the main steel rod^ and hang ano- 
ther pair of steel rods from the top of the brass ones 
to carry the bob at their lower end. Therefore the one 
length of brass has to compensate two lengths of steel ; and 
since the compensating rods must not reach above the 
point of suspension of the pendulum, it will be evident 
tiiat this cannot be done with only one pair of compen- 
sating rods ; and in fact it can only just be done with two 
pairs; for suppose the brass rods to be quite as long as the 
main steel rods, or all the rods to be of the length /, we 
must have iln the expansion of the brass rods upwards 
= Slm the expansion of the steel rods downwards; 
and since — = -61, this is only just possible; and if iron 
is used instead of steel, it is not possible, since the ratio of 
the expansion of iron and brass is *7, which is more 
than 7. And therefore a completely compensated gridiron 
pendulum of steel and brass requires nine bars (as the com- 
|)ensating rods must go in two pairs), and one of iron and 
brass could not be made with less than thirteen bars. Grid- 
iron pendulums have now been superseded by those which 
I shall next describe. 

63. The greats expansion of zinc than brass obviates the 
necessity for so many bars, the ratio of iron and zinc ex- 
pansion being only '41 ; and a very good and elegant and 
tolerably cheap compensated pendulum can be made of iron 
and zinc rods ; or, what is the more simple and common form 
of it, a zinc tube may be made to rest on the regulating nut 


at the bottom of the main rod, and this zinc tube carries an 
iron tube fastened to it at the top, and carrying thq bob at the 
bottom. In compensation by rods it is necessary to add 
more than ^ to the length of zinc aad brass given by the 
calculation in § 60, because the difference between the cen- 
tre of gravity of the bob and the centre of oscillation of 
t)ie pendulum is greater where there are compensating rods 
or tubes of considerable weight, than where the compensa- 
tion is contained in the bob : in other words, the bob ha^ 
to be lower to produce a pendulum equivalent to the re* 
quired simple pendulum when the rod is heavy than when 
it is light. I find that for a l^sec. pendulum (88 inches), 
for which the zinc compensation, if the centres of gravity 
and oscillation were identical, would be 61*6 inches, it is 
necessary to add nearly i to complete the compensation, 
taking into account the weight of the rod and tubes, the bob. 
being a cast iron cylinder a foot long; and the total lengtk 
of the pendulum, from the top of the spring to the bottom, 
of the bob, requires to be nearly 97 inches instead of 88. 

64. There is another kind of compensation, a compound 
of both the former kinds, which was invented by Smeaton^ 
and was generally used by Holmes, a celebrated clockmaker 
of the last century. The rod is of glass, and it carries on 
a collar at the bottom a zinc tube about 12^ inches long, 
from which is hung an iron tube, which carries a lead 
cylindrical bob of the same length as the tubes themselves,^ 
and enclosing them, so that the pendulum looks as if it 
had merely a glass and lead compensation. I wonder it is 
not more frequently made now that glass is cheap, as it 
requires no polishing as a zinc and steel pendulum does. 


or is thought to do^ and I suppose it is equally effective. 
It is evident that the expansion of the zinc and half the 
length of the lead upwards has to compensate that of the 
glass and iron downwards. 

65. In all cases a little additional compensation is re- 
quired on account of the variation in the strength of the pen- 
dulum springs which diminishes as its temperature iucreases. 
The amount of it depends upon the stiffiiess of the spring 
with reference to the weight of the bob and the length of the 
pendulum, and can only be ascertained by trial. According 
to some experiments published by Mr. Dent, the compen- 
sation of the spring requires about ^th to be added to the 
ordinary compensation of the steel rod in a seconds pendulum 
of the usual weight; and consequently about half must be 
added to the compensation required for a wooden rod. I 
am told that in a short pendulum the compensation re- 
quired for the spring was a great deal more than \ of that 
required for the steel rod; and indeed it is evident that this 
must be the case, for the elasticity of the spring must pro- 
duce less effect upon the rate of a long pendulum than a 
short one of the same weight; and therefore in a 14 or even 
an 8 feet pendulum the compensation required for the spring 
must be very little indeed compared with that of the rod. 

66. It is evidently a considerable advantage of a mer- 
curial compensation that it affords the means of actually 
tiying the pendulum and adjusting it so as to compensate 
accurately both the variation of the strength of the spring 
and the expansion of the rod at any two extremes of heat, 
by merely diminishing or increasing the quantity of mer- 
cury ; which cannot be done without a good deal of trouble 


in a pendulum made of solid metals. With the view of 
obtaining a cheaper mercurial compensation than that of a 
cylinder full of that metal I have suggested the following 
method^ which as far as I know is new. Make a koUow 
cast iron ball with a short neck in which a piston fits very 
accurately. It is evident that if the ball is neariiy fiUed 
with mercury and the piston can be made to fit mercury, 
tight, the expansion of the mercury will raise the piston 
through a height which depends upon the ratio of the 
thickness of the piston to the quantity of mercury which 
the ball will contain. Consequently if the ball is fixed to 
the bottom of the pendulum rod by a thing like a pump 
sucker, and the piston is made to carry the bob resting on 
a short cross bar, this apparatus may be used to compensate 
the pendulum, and will require only a small quantity id 
mercury. And as the depth of the piston in the ball does 
not signify, the quantity of mercury can be varied at 
pleasure ; or it may be altered by putting bits of iron into 
the ball in the place of so much mercury. The only diffi- 
culty is to make the piston mercury-tight. Mr. Dent has 
succeeded in doing this for a 39 inch pendulum, even with 
such a heavy bob as 70 lbs. I doubt however whether it 
would be safe to use this plan where the mercury is con- 
stantly under such a heavy pressure as that ; but perhaps it 
may answer for pendulums of twice the weight of those 
which are usually put to astronomical clocks at present, 
and of which the mercury alone costs £4 or £5. 

67. There are two other kinds of compensated pendu-> 
lums which have this peculiarity, that tRey do not depend 
upon the difference of the rates of expansion of their nm^ 


terials. The first of tiiem, the lever compensatioii, vas it 
seems tried Itj Graham, the inventor of the dead escape- 
ment, bnt abandoned by him for the very superior merca- 
lial compensation ; and was afterwards completed by EUicott. 
I have lately seen it nsed in some small Pr^ch clocks, and 
fLerefore it may be worth describing. 

AG is the main rod of iron, 
B D a lever resting on a fulcrum at 
C fix«d to &ie bottom of the rod. 
(There is anoth^. lever, exactly 
the same, only set the opposite 
way, which is with a portion of i 
the rod on the left hand omitted 
in the drawing for greater clear- 
ness). Aba strong collar fixed to 
the main rod anywhere near the 
top. Between this collar and the 
^ort end of each lever ib put a 
bar, either of brass or iron, and 
tlie bob is supported by pins or 
rollers D resting on the long ends 
of the two levers. Now if the 

ratio of the long arm to the short arm of each lever is the 
name as that of the funonnt of expansion of the rod AB to 
that of the main rod, it is evident that the bob will be kept 
in the same place. Theoretically A may be anywhov, and 
AB may be either of iron or of brass; but the less ex- 
pansion AB has the shorter must be the short arm of the 
lever, which is objectionable for several obvious reasons. 
A more serious objection to the pendulum altogether is 


that/ on account of the Mction at D, which ia sliding 
friction of a bad kind, the bob was found to move by 
jumps, and moreover the compensating rods must be very 
thick to avoid bending. It has indeed been proposed to 
remedy this by supporting the greater part of the weight of 
the bob on a spring, and leaving only just enough weight, 
0^ the levers to ke)e|i ti^m to their bearing; but it is com- 
plicated enough lihr^iidljf, and can have no advantage over the 
n^ore simple zim ^OHipensation, until the laws of nature 
are altered, mi itt W&isia expand equally. 

68. Tiiid otW kbd of homogeneous compensation is, 
I think, but foir one. practical difficulty, much more pro- 
mising ; for if this difficulty can be got over, it would 
afford the means of compensating a pendulum 50 feet long 
and a ton in weight, with as little trouble and expense 
as a small one, as well as a very convenient method of ma- 
king the finer adjustments for time, without stopping the 
pendulum. If the pendulum spring, instead of bendiog 
^m a fixed poiat, is passed through a sUt ui a fixed cock, 
and is itself carried by another cock above, which admits of 
being raised as much as the pendulum increases in length, 
the effective length of the pendulum will evidently remain 
the same. Therefore if the upper cock is made to slide in 
a vertical grove and rests upon an upright bar of the same 
material and length as the pendulum, standing on a firm 
support at the bottom, the expansion of this bar upwards 
will exactly compensate the expansion of the pendulum rod 
downwards. And in favour of this method it is said that 
if the bar and the pendulum rod are taken from the same, 
piece, we are sure that their rate of expansion will be the 



same, whereas no two experiments witK different pieces of 
the same metal ^ve exactly the same result. But as this 
rod acts by poshing, not by pulling, it would have to be 
very thick to support the cock for a long and a heavy pen- 
dolnm without bending. 

69. The sune object might however be effected in ano- 
ther way and with sevoral incidental advantages, provided 
the objection I shall mention can be removed. I«t tiie 
pendulum be hung from a fixed cock A as usual, only a few 
inches higher, and with tlie spring reaching a few inches 
above the end of the pallet arbor. B C D is a cast iron 
lever with its pi- 
vots at C turning 
in two Vs in a 
strong cock firmly 
fixed to the wall. 
At one end D is 
fixed a wire or rod 
SEFof the same 
Ifaigth and mate- 
rial as the pendu- 
lum rod, A squa- 
red portion of this 
rod passes through 
the cock E to keep 
it from twisting, 
and ends in ascrew 
passing throng 
another coekF, and 
has a nut G below. 


E and F would be cast all in one piece and fixed to the wall* 
The other end B of the lever at the same distance as D is from G 

is so made as to clip the pendulum spring, but not so tightly 
that it cannot slide under a moderate pressure ; and in order 
to produce this pressure and keep the lever steady, a weight 
W is hung on to that end of the lever. Then as the pen- 
dulum rod lengthens, the compensating rod lengthens too, 
and lets the weight W pull down the end B of the lever, 
and so makes the effective length of the pendulum the same 
as before. As the motion would not have to be quite -^ of 
an inch to compensate a 14 feet pendulum for 40^ of heat, it is 
not of the least <»nsequence that the end of the lever moves 
in a circular arc and not in a straight line; and the longer the 
pendulum is the longer the lever should be in order to dimin* 
ish the curvature of the arc which its end has to describe. . 
The practical objection to this kind of compensation, to 
which I referred, is, that any alteration in the length of the 
spring is found to affect the rate of the pendulum, not uni- 
formly, but in some variable way which is not reducible to any 
fixed law. This is entiidy an experimental question, and I 
have no means of knowing to what extent this variation takes 
place, or whether any experiment exactly similar to this has 
been tried. I do not suppose that this compensation would 
be equal to the others; but considering the advantages of long 
and heavy compensated pendulums over short or uncompen- 
sated ones, tEai the impossibility of compensating very long 
pendulums at a moderate expense by any of the usual methods, 
I should be inclined to try it where the building affords &cil- 
ities for a pendulum as long as 21 feet (2^ sec), such as there 
is in Doncaster Church, with however ft very inferior dock. 


quite iinfit for tke work it has to do. I may add that when 
the tower shakes under the ringing of the bells^ as all 
towers that have not spires on them do with a heavy peal of 
bells^ a long and heavy pendulum is especially necessary.* 

70. There are a few other matters relating to pendulums 
which may as well be mentioned here. First as to their 
shape. Until lately pendulum bobs^ except mercurial ones^ 
were almost universally of the shape of a common magnifyr 
ing glass or lens, and such pendulums are therefore called 
lenticular. This sliape was adopted on account of its pas- 
sing through the air with less resistance than any other 
figure that could conveniently be used. But they are liable 
to this objection : if one of these bobs were set on a rod 
with its broad side instead of its edge towards the directioDi 
of motion, the pendulimi would vibiiate in a different time 
(independently of the resistance of the air), because the 
moment of inertia of a circular plate turning round an axis 
placed anywhere at right angles to its plane is greater than 
if the same axis were placed in the plane. Consequently if 
the pendulum rod should from any cause get so placed or 
twisted that the lenticular bob (which, approaches to a f^t 
plate) does not always vibrate with its largest or central 
circle in the plane of motion of the pendulum rod, the time 

* I take thb oppoTtanity of correcting a mistake in Southey's 
well-known book called The Doctor, wherein he laments over the re- 
moval of the old peal of bells in Doncaster Church on account of the 
inability of the tower to bear them any longer. The old peal mr^ 
removed in 1835, but only to be recast, the two largest beUs being 
cracked ; and the new peal was put up in the same year, and is one 
of the heaviest peals of eight in the kingdom, the tenor weighing 


bf vibration will vary; and in nearly every common clock 
the bob of the pendulmn may be observed to have that kind 
of twisting motion which is familiarly termed. wMUng, es- 
pecially when it swings a large arc. 

But this twist of the bob can produce no effect upon 
the time of vibration, if the bob is a cylinder or any other 
solid of revolution having the rod of the pendulum for its 
axis. And the pendulum is also less likely to acquire that 
motion, if the bob is such a solid of revolution, than if it 
presents a surface unequally exposed to the resistance of the 
air, as the lenticular bob does if it is ever so little twisted. 
A sphere (which Holmes put by Smeaton's advice to the 
Greenwich Hospital clock) is iaconvenient, not only because 
of its requiring so much width for the pendulum to swing 
in, but because a slight error in making the hole for the rod, 
which throws it out of the axis of the sphere, causes a 
greater disproportion between the mass on each side of the 
rod than in a longish cylinder or a prolate spheriod having 
the rod for its axis. Therefore upon the whole a cylinder 
is probably the best shape for the bob ; and there is no 
objection to its top being made a portion of a sphere, when 
the bob is made of cast iron, as it rather improves the ap- 
pearance, and also prevents anything from settling upon the 
top, as sometimes happens in church clocks, with bits of 
mortar and the like, the effect of which we shall see pre- 
sently. The pendulum of the clock in the frontispiece is 
of this shape. It may be convenient to state that a casrt 
iron cylinder 9 inches wide and 12 inches high, with a 
spherical'top rising 2 inches more, and a hole 2 inches wide 
all through it for the compensation tubes, weighs 2 cwt. 


within a very few pounds ; and the centre of gravity of the 
bob is about 7i inches from the top. 

71. The next point is the regulation of the p^dulum* 
The common and well-known method is to raise the bob 
by a screw at the bottom of the rod when the dock loses^ 
and to lower it when it gains ; and the quantity by which 
this must be done for any given bss or gain of the dock 
may be determined approximatdy in the manner stated in 
the note at page 83« But it is difficult to do this without 
stopping, or least disturbing the pendulum, which requires 
the dock to be set again ; and therefore two other methods 
are used for making the finer adjustments in regulators and 
in turret clocks with heavy pendulums, which require a strong 
and coarse thread to the screw, and must be hdd steady while 
the screw is turned to avoid twisting the pendulum spring. 
One is, to have a small weight sliding on the rod and kept 
in its place by a spring collar, which allows it to be moved 
up and down by a blow with a small hammer; raising it 
accderates the pendulum, only in a much smaller de- 
gree than raising the. heavy bob, and therefore it admits 
of greater accuracy, The other method is to have a few 
small weights to lay on the top of .the bob ; adding th^n 
of course raises the centre of gravity, and of oscillation, 
and accelerates the pendulum; if they were put si the 
bottom of the bob they would produce the contrary effect. 
When the bob has a round top the little weights may be set 
on spikes (of whi^h there should be a pair, one on each side 
of the rod)/ or in a cup, or on a ring on the top of the bob. 

The amount of accderation produced by any given 

small weight put on the top o! the bob, will evidently vary 



with the length of the hoh directly (that is with the height 
above the centre of gravity or oscillation at which the little 
weight is placed), and the length of the rod, and the weight 
of the bob, iwvenely. Let / as before be the length from 
the top of the pendnlnm to the c.g.iA the bob (assuming it 
as before to be identical with the centre of oscillation), h 
the distance of the e.g. of the bob from its top, M the 
weight of the bob, m the little weight to be added, and dl 
the qnantily that the e.g^ is raised thereby; then it is 
easily proved that ^^=::^~r- = — ^if wepntr for — • And 
as in § 57, the corresponding daily acceleration d'L will 
= 7-TTw itt seconds. Suppose, for example, that -j = -^ 
(as if / is 88 inches, a 1^ second pendulum, and b^=1\ in.), 
M = 2 cwt., andfl» = lib.; then^^T = jg§^= 16 sec; 
or an ounce weight put on the bob of such a pendulum 
would accelerate it just one second a day. The pendulum 
of the Exchange clock, a 2 seconds pendulum with a bob 20 
inches long and weighing above 3 cwt., was regulated sojne 
time ago by putting a penny on it. 

We may now leave the pendulum and escapement, 
which may be considered the mathematical parts of a clock, 
and return to the consideration of the merely mechanical 


72. The general construction of the going or time« 
keeping part has been explained already (18). But for 
docks that are to keep accurate time, there is yet something 
wanting, viz., a contrivance to keep them going while they 



are bemg wound np, for that of conise takes the action 
of the weight off the dock, and so the scape-wheel will not 
escape nnlesa some equivalent piessnre ia sppUed to the tiain. 
In common house-clocks and the cheapest watches there is no 
Buch contriTance, and they stand still while they are wind- 
ing np ; bat in all others the apparatus which is applied for 
the purpose is that invented by Harrison, and called the 
ffoitif barrel, or rtdehett, in we^t clocks, uid the ff<mg 
fiuee in watches and spring clocks. This drawing shows 
the arrangement of 
it. The barrel with 
its ratchett and 
click are the se 
as before 
plained; bnt the 
dick c is not now 
placedon the great 
wheel, but upon 
. another whed rid- 
ing on the arbor 
of the barrel be- 
tween the great 
whed and the barrd ratchett : this whed has also ratchett 
teeth cot upon it, but turned the opposite way to the barrel 
latdiett, as shown at B, and the dick EC bdonging to it is 
s long arm turning on a pivot C in the dock-&ame ; and 
this second ratchett whed is connected with the great wheel 
in any convenient way by a spring S S, having oiLe end Sxed 
OD the ratdiett wheel, and the other end pressing against a 
pta on the great wheeL The action is this : while the 
7 2 

100 CLOCKS^ 

dock is going, the weight polls the barrel and both 
ratchetts to the left, and the going ratchett, by means of 
the spring S S, presses the great wheel in the same direO' 
tion ; and as the clock goes on, one tooth after another 
of the ratchett R slips nnder the long dick, and this cansea 
the drop wUch may be heard about every five minut(» 
in regolators and in good watches. When the weight 
is taken off the barrel by winding up, the going ratchett 
immediately flies back a Uttle towards the right, but is 
stopped as soon as one of the teeth arrives at the cUck, and 
there it is held ; the spring continues to press the great 
wheel as before, with nearly as much force as when the 
weight is acting, and so keeps the wheel in motion for this 
short time that the clock is winding up. 


73. The only thing that remains to be described in the. 
going part of clocks is the dial work, or the machinery by 
which the hands are moved. The minute hand is set upon 
the square end of a socket or pipe, which fits rather tightly 
ou the long projecting arbor of the centre wheel. It must 
not be quite tight, or the hand could not be put forward or 
backward when the clock wants altering ; and the requisite 
degree of tightness or friction is obtained as follows. 
The pipe on which the hand is set is about an inch long, 
and has a whed fixed to its other end (the use of which will 
be described in the next section, and which may be called 
the hour<-whed), and the pipe slides on to the arbor pretty 
easily ; but before it is put on, a slightly bent oval spring, 
nearly as bug as the diameter of the whed, with a squife 


hole in thie middle/ is dipped on to tlie arbpr with the 
concave side outwards^ and the square hole fitting on 
to a corresponding square, cut for about ^th of an 
inch on the arbor, so that .the spring is always car- 
ried round with the arbor; then the pipe is put on, and 
the wheel rests against the concave ^eof the spring ; then 
(omitting the hour-hand for the present) the hand is put 
on, and generally a small cap or collar in front of it; the 
collar is pushed back so as to force the wheel more tightly 
against the spring, and a pin is, put through the end of the 
arbor just in front of the collar. It is evident, therefore, 
that the hand and its pipe will be kept steady on the arbor 
by the pressure between, the collar and the spring, and the 
friction of the spring upon the hour wheel, but that it can 
be turned when necessary. Sometimes the hole in the 
spring is left round instead of square because it is less trou- 
ble, but it is a slovenly practice, and it requires much more 
pressure to produce the same degree of steadiness when the 
friction of the spring acts upon the small diameter of the arbor 
than when it acts upon the large diameter of the hour wheel, 
74^ Besides carrying its own hand, the long hand pipe, 
and the wheel belonging to it, have to turn the short hand, 
and also in striking clocks to let off the striking part. The 
form of hour-hand which was formerly used in regulators is 
this : the minute-hand socket has a pinion on it, and this 
pinion works a large .wheel with twelve times as many teeth 
as the pinion, which therefore turns round once in twelve 
hours ; this wheel has the twelve hours engraved upon it, and 
there is a hole in the dial though which the figures appear, 
and a stationary hand or index pointing to the proper figure. 

102 CLOCKS. 

This might have been done in a still more simple way> 
and without incuiring the friction of that wheel and pinion, 
by setting the hour drde on the arbor, of the great wheel, 
which turns in twelve hours. It is of no consequence that 
it would be turned with the arbor in winding up, because 
you would only have to stop winding at a place which would 
leave the proper figure of the hour circle under the index. 

These moveable hour circles are now abandoned^ on 
account of the small figures moving past a hole being so 
much less easy to see than a hand moving in the usual way. 
The now obsolete day-of-the-month wheels, which required 
setting for five months in the year, showed themselves 
through a hole in the same way. But if a hand were 
merely put in the place of the moveable dial, it would turn 
the wrong way, that is the opposite way to the other hand | 
and therefore an intermediate wheel is now put betwe^ the 
pinion of the minute-hand socket and the hour-hand wheel, 
and the multiplier 12 (when there is no striking) may be 
divided between these three wheels and their pinions as we 
please ; and in regulators the hour-hand usually points to a 
separate dial, or set of figures engraved on the lower part of 
the large dial, and corresponding to the small dial for the 
seconds-hand on the upper part. 

There is another way, which is sometimes adopted, of 
working the short hand on a separate dial, and without the 
intervention of this second wheel, viz., by putting the hand 
upon an axis which goes through the frame of the clock, 
and carries a wheel working into another equal wheel fixed 
on to the great wheel, which turns in twelve hours. The 
great wheel itself is so large that it would be inconvenient 


to have another of the same size working into it, and thii 
is the reason of this other wheel being fixed to it. There 
is so much less Motion in this way of working the hour 
hand directly from the great wheel than by going up to the 
centre wheel and then down again by two more wheeb, that 
I wonder it is not more commonly adopted. The only 
objection to it is that if the hands want setting they must 
be set separately; but a regulator dock requires so little 
alteration that the hour hand never has to be meddled with 
unless the dock has been allowed to stop. 

75. AH these methods howeyer are different from that 
which must be used when both the hands turn on an axis 
in the middle of the dial, as they do in all clocks ezcept 
astronomical ones, and sometimes in them. In that case 
there is a wheel of the same size as the whed we have 
called the hour whed placed by the side of it and driven 
by it, and therefore turning the toronff way once in an hour, 
and having on its arbor a pinion with six or more leaves : 
this whed may be called the reverse hour wheeL Its 
pinion drives a whed with twdve times as many teeth, 
which is fixed to a socket or pipe riding on the former one ; 
and this pipe carries the hour hand, and will evidently turn 
in the rigit direction once in twdve hours. This hour 
hand pipe is not indeed really carried by the minute hand 
one ; for in order to take the wdght of it off the centre 
whed arbor, it rides upon another fixed socket or pipe sur- 
rounding the minute hand pipe and set upon a cock which 
extends over the minute hand whed and is screwed to the 
frame: this fixed cock and pipe are called ^ cannon^ So 
that the apparent central axis of the hands of a common 

104 CLOCKS. 

dock oonaiflts, first, of the centre wheel arbor; seeondljr', 
of the pipe to which the long hand is fixed; thirdly, of the 
cannon; and fourthly, of the hour hand pipe; and, fifthly^ 
if there is an alarum to the clock, of the pipe to which the 
little alarum dial is. fixed* This arrang^n^t of the dial 
woik is the.most cinmsy part of Eng^h and French dbdcs. 
It is done better in the American; and I shall propose 
what I think a stQl better arrangement ofit, in §87, in con<^ 
nexion with another alteration of the usual construction of 
eight-day docks.. Occasionally the seconds hand is also 
set upon a thin arbor within all the others and traveb 
round the large dial, instead of having a small one to 
itaelf ; this is a very bad practice on account of the in* 
creased friction, which it causes, and such docks are rarely 
made now. 


76. The thingiit called Equation Clocks bdong to the 
subject of dial work; for they are (or rather were) docks 
for showings but not going, true instead of mean solar 
time; which was done by communicating a secondary or 
superimposed motion to the hands, so as to advance or 
retard them according to the equation of time for every 
day in the year, lliey were never in common use in 
Bngland, though they have occasionally been made as 
curiosities; but Mr. Yulliamy says in his pamphlet before 
quoted, that until December, 1826, 'all the French public 
docks of any repute were made to show solar time.' The 
machinery by which it was done was, as may be supposed^ 
complicated; and as it is perfectly useless, and worse, than 


useless^ and a description of it may be found in Said's 
Treatise on €loekmaking and other books^ I shall not 
describe it. For house clocks there appears to have been 
a somewhat more simple method of doing it^ by making the 
dial itself revolve slowly^ according to the equation of time^ 
while the hands went uniformly in the usual way. The 
dial would have to turn about half round between November 
and February, the figure XII appearing nearly at the usual 
place of IX in November, and at III in February, with 
other smaUer oscillations in t}ie remaining nine months. 
And after all the ingenuity thrown away upon these con-^ 
trivances, true time could not be shown by any of them as 
accurately as by a common clock, with an equation table 
such as that printed at page 11. For the equational motion 
was, in all the plans, communicated to the dial or dial work 
by means of a plate of an irregular oval form set upon a 
wheel revolving in a year; the form of the plate was such 
that the radii, from the axis on which it turned to the 
edge of the plate, were proportionate to the quantity by 
which the clock ought to be before or behind the sun; and 
tiie plate as it revolved pushed some moving frame work 
jEEurtber from the centre or nearer according to the time of 
year* And it is evident that no great accuracy could be ob-r 
tained in this way except upon an inconveniently large scale^ 


77. Occasionally clocks are made to go for a month or 

even a year. This is done merely by the addition of one 

wheel or two below the usual great wheel, and increasing 

the weight to four or forty-six times as much as usual, 

p a 

106 CLOCKS. 

Occasionally two barrels are naeA, to avoid the great straiii 
upon the teeth of one great wheel and the adjacent pinion. 
One string is then sometimes made to mn off both baiteb 
together^ carrying one weight between the two lines ; which 
however has a pulley for the string to pass through to 
keep its tension equals though the pulley has no sensible 
motion. The barrels will of course turn half as many 
times round in the month w year as if one end of eadi 
string was fixed to the frame in the usual way^ and the one 
weight was divided into two. Chronometers are also made 
to go long periods in a similar way^ by the use of two or 
more main-springs to drive one wheels as tins secures 
greater uniformity of force than one very strong spring* 


78. These clocks are seldom made now as English or 
French house clocks. Most of the Dutch and many of the 
American clocks however are so. When a thirty-hour 
clock winds up with a key like other clocks, in order that 
it may have only three wheels in the train, the minute 
hand arbor does not belong to a wheel in the tnun, but to 
a supernumerary wheel, which is driven by the great wheel 
in the same way as in one of the methods I described for 
working the hour hand' of an astronomical clock, and the 
great wheel turns in two or three hours. In the Dutch 
clocks and the old English thirty-hour clocks, the arbor 
of the great wheel is not used to wind up by, but comes 
through the frame and carries a wheel which works the 
hour and minute hand wheels both together, being smaller 
than one and largw than the other : or some equivalent 


anangemeot, of which many vadetiee mi^ be conoeived. 
Fot as the arbor of the great wheel is used to cany tliiB 
dial working wheel, it evidently canuot be used also to 
wind up the barrel; uid accordii^ly instead of a barrel 
there is used a sort of dee^ pulley with spikes in it to 
prevfait the rope or chain from slipping : this pulley is not 
fixed to the aibor but lidea upon it, and it is connected 
with the great wheel by a ratchett and click as bef(»e des- 
cribed* One end of the rope carries the clock weight loid 
the other end a little w«i^t merely to keep the rope steady 
and to take hold of to paU up the great weight by when 
you wind np the clock. 

7 9. There is however another way of appl^ug this sort of 
spiked pnlley, which 
deserves noticeon bc- 
conntof itsprodudng 
the cnrioos effect of 
keqiing the weight 
sdingupoa the dock, 
at the same time that 
it is being wound up. 
Itis called the endlett 
eioM of Hnygens. 
P is the spiked pulley 
fixed on to the great 
whed Bibor, not now 
by a ratehett and 
click but rigidly ; p 
whichmay rideeil^ 

108 CLOCKS. 

•on the same axis or on any other conveniently placed, and 
having a latchett and a cUck fixed to the clock frame. The 
midless chain abed passes over both pnlleys, and in the 
loop formed by c^ on the right side of both pnlleys it car- 
ries the weight W hnng by a common pulley; a little 
counterbalancing weight to is hung to the other loop ab ; 
this little weight is somelames merely a large ring hung on 
to the chain. Now if the string b is pulled down by hand 
it will puU up c and so raise the weight W, which being 
hung by a moveable pulley will nevertheless still press upon 
the string d and so keep the elock in motion by pressing 
on the right side of the principal pulley P. If the clock 
ha3 a striking part, p may be the pulley of the great wheel 
of that part, the click behig fixed to that wheel; and then 
the clock may be wound up at any time when it is not 
striking. Only in that case it must be remembered that 
rWght wiU descend in half the time that it descend, in 
when the pulley p never has any motion to the right, or 
the click is fixed to the frame ; and the same will be the 
case with only a going part, if the cUck is fixed to the 
great wheel instead of to the clock &ame, for then both 
puUeys will turn together ; but while the clock is being 
wound up only half the weight will then act upon the great 



80. There is yet another kind of elock, which difiers 
from all that I have yet described in having for its main- 
taining power not a weight but a long spiral spring, which 
is coiled up in a box or barrel as tightly as possible when 
it is wound up, but with room to uncoil itself for a few 


toms of the bairelj in doing which it moves the train 
which is connected with the barrel just as a weight does. 
The simplest constroctioii^ which is used in all the French 
clocks and watches^ is this : the great wheel is fixed to the 
barrel that contains the spring; the outer end of the spring 
being fastened to the barrel^ and the inner end to a strong 
arbor which goes through the barrel, turning in the caps of 
the barrel, and ends in the winding square, and also carries 
a ratchett wheel, of which the click is fixed to the dock 
frame, to hold it as you wind up. And therefore this 
kind of clock requires no maintaining power to keep it 
going while winding up; for the tension of the spring is 
acting upon the barrel at that time as much as any other, 
in fact rather more. But on the other hand the force of 
the sprinff is much crreater when it has been just wound 
up, tJaTwhen it is^ly run down; with however this 
lemarkable, and apparently anomalons exception j that it is 
found that there is a certain position of the spring in 
which its force is nearly the same for three or four turns, 
but no more. Consequently, as a watch only requires a 
few turns to wind it up for a day, the variation in force in 
!French watches is slight, compared with that in the clocks 
which go a week or a fortnight, and in which the spring 
has many more turns to make than in a watch* This is 
one reason of the inferiority of French watches, and still more 
80 of clocks^ to English ones, both in price and quality. . 

81. For this inequality of force is removed in English 

spring clocks and watches by the use of what is called the 

fmee. The fasee is nearly a cone, that is to say, a cone 

whose slant side is concave instead of straight, with a spiral 

110 CLOCKS. 

groove cut round it from one end to. the other, as on the 
barrel of a weight-dock, for a string or chain to ran in* 
The spring is enclosed in a barrel as before, but the barrel 
does not carry a wheel, and its arbor, instead of being used 
to wind up by, has a ratchett and click merdj for the 
purpose of adjusting the tension of the spring. A catgut 
cord, or a chain, is wound round the barrel, and the loose 
end of it fastened to the thick end of the fusee. The fosee 
is fixed upon the winding arbor, and has the great wheel 
riding upon it with a ratchett, just as the great wheel of a 
weight-clock rides on the barrel arbor ; and the best clocks 
and watches have a 'going ratchett' also. As the fus^ is 
wound up, it draws the cord off the barrel along the groove 
until it is all wound off the spring barrel on to the fusee* 
And it is evident that when the tension of the spring is 
greatest, it is pulling at the thin end of the fusee, and when 
the tension is least it puUs at the thick end ; and therefore, 
if the radius of the fusee is made to decrease, as the force of 
the sprii^ at the corresponding point of winding up in^ 
creases, the force communicated to the great wheel will be 

S2,, In order to prevent the cord from being broken by 
over-winding, there i3 a kind of lever fixed to the frame;^ 
with a hook to it, so placed that as the cord advances to* 
wards the thin end of the fusee, it pushes this lever aside, 
so. that its hook catches hold of a long tooth projecting from 
the end or cap of the fosee, and stops it from being, wound 
up any farther. As the cord recedes agaia, the lever is 
allowed to recede under the pressure of a small spring behind 
it, and by the time the fusee has made one revolution^ the 


lever and its hook have got oat of the vay of t]ie long 

83. Ftobabljr few readers of this book require to be told 
what the chain of a watch or a spring clock is like, bat it is 
proper to describe it in a ' radimentar^ treatise.' It coa- 
msta, then, not of links set across each other, bat of pistes, 
alternately' one and two, rivetted together; bat the hole in 
the single plates is bo large that the rivet does not stick 
tightly ia it. I 

Here are two I 
pairs of links,! 
drawn on a scale I 
wEiich the reader I 

will think more St for a drawing of the chains of a suspen- 
sion bridge than of a watch or even a clock. A. chain of 
this kind never twists, but will only bend in a plane (co- 
nearly so) parallel to the plates of which it is composed. 

84, These clocks are most freqnently made to stand on 
a bracket, and alwa^ of snch a size that their pendnlnms 
can only be abont 10 or 18 inches long; and as their bobs 
are small also, we know what the resalt most be as regards 
their accaracy. However, when they are well made, with a 
fosee, and not exposed to a temperature which freezes the oil 
(whic^ is much above the freezing point of water), they will 
go nearly as well as a coarsely made long clock of the old 
fiishioned kind. Sometimes they require a good deal of 
trouble to set them so as to beat equally, for if they are not 
■0 set, they are very likely to stop, as they have g^ierally, 
and the foreign ones always, very httle force to spare. This 
position is not, as is conmumly suj^sed, that of actual ver- 

112 CLOCKS, 

iicalitj, but merdj of verticalitj relatively to the escape- 
ment ; they ought indeed to coincide^ but they frequently do 
not; and oraisequently after the bracket has been made 
quite level, it is found either that the crutch wants bending^ 
or the clock raising on one side by bits of card. They are 
much more likely to stand steady, and also easier to adjust, 
if the two hind feet are taken off, and one put instead in the 
middle between where they were. 

85, These are now almost the only. English clocks 
(except regulators) that find any sale; and even these are 
getting fast superseded by the better class of American 
clocks, and by l^nch ornamental clocks, neither of which, 
however, will last nearly so long. With the latter it is no 
doubt quite hopeless for us to compete, as, besides the 
greater cheapness of their labour, the French appear to pos- 
sess what I may call a smaller eye and finger than English 
workmen, and they are able to perform delicate and oma* 
mental work with much greater quickness and facility. And 
as those who chiefly regard the beauty of the figure of their 
clocks seldom care much about their entrails, they consider it 
of no consequence that a good English clock is better for the 
natural object of a clock than a foreign one. Whether it 
would be possible to mcmufactwre clocks on a large scale as 
cheap as the American ones, I am not able to judge. I have 
been told that, but for the cases, it would. But unless the 
English dockmakers take some steps towards either altering 
the kind of docks that they make, or can fiind out some 
cheaper mode of making them, there is no doubt that there 
wiU soon be no house clocks, except regulators, made in 
this country. The old-fashioned, fiiU-length house clock i^ 


now nearly exploded^ on account of its ugliness^ size, and 
deamess, as compared with the American clocks, which go 
Bofficientlj well for ordinary purchasers. 

No one who has seen the inside of an American clock 
€an help seeing that onrs are unnecessarily heavy, and waste 
a great deal of the force in merely overcoming their own 
inertia and friction. An American clock goes a week with 
both the weight and the fall for it not half of what they ai6 
in the common English clocks ; and as a large pendulum 
requires no more force to keep it going than a small one, it 
is evident that about fths of the moving power in our 
clocks is wasted. (The commendation of the American 
docks cannot be extended to the fixing of their pendulums, 
which is as bad as possible.) I have also seen some very neat 
French clocks, about the same size as the American, but 
much more highly finished, and with dead escapements, 
going a week with a very small weight. 

86. A clock might be made very much better than 
any English spring dock, or any of the foreign ones, and 
quite as cheaply as the old long clock, and at least equal to 
it in performance and very superior in appearance, and 
therefore more likely to sell, by making the case only of the 
length required for a seconds pendulum. This may be done 
in three ways : 1. By hanging the weights by three strings 
instead of two : the lower pulley should be let into the 
weight instead of occupying the height of 3 or 4 inches 
above it, and should be as broad as the space between the 
weights will allow. I have converted an old long clock 
into one of 'three-quarter length' in this way, and cut off 
the large and ugly and useless pedestal of the case. The 

114 CLOCKS. 

second method is to iftake the barrel ids of the usual diame- 
ter; and the third and best is to make the number of teeth 
in the great wheel half as many again. Or the second and 
third methods may be combined in any convenient propor- 
tion. In all these cases of course the weight will have to 
be half as large again, except that, as I said before, it in 
now unnecessarily heavy, because the wheels are so. I may 
observe also that the third of the above methods will re- 
quire a barrel of only |ds of the usual length, and therefore 
all the arbors and the space between the plates may be so 
much less. Mr. Dent now makes nearly all his regulators 
■*^ in this way, and such clocks take up no more room, except 
in length, than a common spring clock; and for all the 
reasons I have mentioned I beg to suggest' the adoption oi 
this size of clock to both the makers and the purchasers of 
clocks : the faces should be about the same size as those of 
common spring clocks ; one door serves for the body and 
face, or the whole of the front and sides take off at once. 

87. And with this alteration I think there may very 

well be combined that alteration of the dial work which I 

referred to in § 75, as it will also help the arrangement of 

making the barrel to turn only 11 times instead of 16 in 

8 days. Those who are acquainted with the American 

clocks will see that this plan is partly borrowed from them. 

The arrangement will be pretty clear from this drawing. 

The scape wheel and second wheel will be worked by the 

centre or 90 minute wheel as usual, which may have 70 

teeth driving a pinion of 7 on the second wheel with 54 

teeth driving the scape wheel pinion of 6. I adopt these 

small numbers on the supposition that all these driven 


pinions are to be lantern pinioiu, as tliey are in tbe Ame. 
rican clocks, which is one of the reasons of their going 
with such small weights, as will be explained in § 126 'on 
the teeth of wheels.' The pinion of 8 on the left side of 
drawing, not being driven, but driving the 11 horn wheel 
of 64 teeth, is not to be a lantern pinion bat a common 
one, only hollow, as it is fixed to the wheel of 46 which 
rides on the centre wheel arbor and tuns with it in an honr 
and an half, and admits of being moved on the arbor when 
yon alter the hands, being held by the pin and the bent 
spring shown in the drawing, as before described in ^ 73. 



The advantages of this phin are, fiwt that the ftiction 
of the large and heavy 12 .hour wheel pipe riding on the 
' cannon ' is got rid of, or rather that pipe ia reduced in 
size so that it need not be any larger than the usual hour 
wheel pipe; secondly, the minute hand pipe is got rid of 
entirdy, and consequently the weight and friction of that 
arbor is also reduced ; thirdly, the 12 hour wheel ia driven 
directly by the tram instead of in the usual roundabout 
way ; fourthly, the discharging pin is put on the hour wheel 
itself instead of on the usual reversed hour wheel, and con- 
sequently it must always agree with the hand ; whereas at 
present clocks are not unfrequently put together so as to 
strike when the hand is nearly a minute from the proper 
place : thie hour hand can also be adjusted by simply pulling 
its pipe forward when the minute hand is off, as there is 
room to slide tiie 12 hour whed out of gear with 4he pinion 
that drives it. The hour arbor is to be kept in its place by 
a small collar at its end, over which there stands a cock 
which may be screwed on from the outside of the clock 
framie. . The lifting piece of the striking part will have to 
be inside the frame; but that, as any clockmaker will see is 
of no consequence, as it will only require a pin to go 
through the hole in the front plate to lift the click of the 
repeating work, instead of going through from the front to 
the inside to stop the pin wheel as usual. In the striking 
part the reduction of the turns of the barrel from 16 to 11 
may be made (besides the methods above described for the 
going part) by increasing the pins and teeth of the striking 
wheel in the proportion of 8 to 10 and giving the great 
wheel 93 teeth instead of 78, And I have no doubt that 


if a Spiral spring is put to the hammer, as I shal! describe 
presently, and the higher wheels made aa light as the 
American clock wheels, the present size of clock weight 
would atiU be quit« sufficient in the striking as well as the 
going part. However I shall leave this to the consideratioQ 
of clockmakers and proceed to describe the 

88. A clock may be made to etnke one at every honr 
without any separate striking part, fay merely putting a pin 

118 CLOCKS. 

into either of the wheels of the dial work that tarns in an 
hour, and a hammer tail or lever over it^ so that the pin 
will begin to raise the hammer aboat a quarter before the 
honr^ and jnst slip from under it when the minute hand 
points to the hour. Instead of the pin there may be put 
upon the front of the wheel what is called a mail^ which is 
a flat piece of metal cut into a spiral form, as shown in the 
drawing ; the efiect of which is that the work of raising 
the lever is distributed over the whole hour instead of 
having all to be done in a quarter of an hour, but with con- 
siderable friction. I have drawn the bammer as worked by 
a spiral spnng, instead of the usual long and stiff spring, 
which produces a great pressure on the pivot^ and is no 
better than the inside bit of a broken watch spring in any 
respect, and not so cheap to make and fix. The arbor 
should be made thicker than usual at the place where the 
spring is put on. This apparatus will require an addition 
of about one-sixth to the clock weight, as there are 12 blows 
to be struck in the 12 hours instead of 78, and the striking 
and going weight are usually about equal. Clocks of this kind 
are put in some of the C!ourts of Law at Westminster, and are 
just as good for a room as those that strike the number of 
the hours, since they call attention to the fact of the hour 
being up, and any body who does not know what the hour 
must be has only to look. They can of course be made 
cheaper than full striking clocks, which require a separate 
train of wheels, and which are made as follows. 

On one of these wheels are placed any convenient num- 
ber of pins so as to raise the hammer tail in succession. 
In eight-day clocks the pins are put upon the wh^ next 


to the great wheels and are 8 or in coarse clocks 6 ; and 
the pinion liaving as many leaves as pins^ the great wheel will 
turn in the 12 hours if it has 78 teeth^ or one for every hlow 
struck in 12 hours. In thirty-hour clocks the pins are set 
upon the great wheel. Above the wheel which has the strik- 
ing pins there are usually two others^ the highest of which 
drives a fan or fly to regulate the velocity of the train : a 
pendulum and escapement have been proposed for the same 
purpose^ but the fly is perfectly sufficient. One of these two 
wheels has another office to perform besides driving the 
fly. The one above the striking wheel must turn exactly 
round for one or more blows of the hammer. I shall as- 
sume it to be once for every blow as it usually is. 

89. The principle of all the various ways of letting off 
the striking part wiU be seen from the next drawing. A pin 
P in one of the higher wheels of the traia^ called the pin 
wheels rests against a stop S on a lever^ called the lifHn^ 
piece, when the clock is not striking. This liftiug piece 
turns on a pivot and has a tail^ which is raised by a pin R 
or a snail on one of the hour wheels of the going part when 
the dock is within a few minutes of striking^ and this lets 
the pin P slip past the stop S on the lifting piece; but it is 
' not allowed to go far, being presently detained by another 
^ stop T either on the same lifting piece or on another con- 
; neoted *ith it; the noise made by this is caUed giving 
varmn^. When the tune is come for striking^ the pin R 
I slips from under the lifting piece and lets it drop^ and so 
f the pin wheel can turn round until the pin P has come to 8 
* again. But before that happens provision is made for getting 
8 out g[ the way if the clock has to strike more than one. 

This is done as follows : there is a large wheel called the 
loclamg^late or cotint-wieel, which tams in ] 2 hours, and 
may therdore conyeniently be put upon the great wheel. 
The rim of this locking-plate is marked out into 78 divi- 
sions, and then deep notches are cut in it at the BUcoessiTe 
distances of 1, 2,8, &c., of the divisions, up to li,iiewot 
which are shown in the drawing. The lifting piece has ano- 
ther arm which reaches as &r as the locking-plate and has a 
tooth which just fits into the notches, Now while the first 
blow of any how is striking and the pin wheel is makii^ 


one turn, the locking-plate is turning also in the direction 
shown in the drawing ; and in so doing the notch lifts the 
tooth of the lifting piece out and makes it rest on the rim 
until another notch has arrived for it to fall into ; and the 
depth of the notches is such that the lifting piece is moved 
far enough to lift both the stops S and T quite out of the 
way of the pin P by the time it has gone once round, and 
they remain there until the clock has struck the proper 
number and the tooth of the lifting piece drops into the 
next notch in the locking-plate : which is in fact an hour 
dial if the notches are marked with the hours, as the num- 
ber under the lifting-piece is always the last number the 
dock has struck. 

Between twelve and two o^clock it will be seen on looking 
at a locking-plate there is one long cut instead of two nicks, 
as I have shown in the drawing, because the clock will strike 
one blow without having the lifting-piece lifted at all, except 
as the going part lifts it. And in like manner in the French 
spring clocks which are often made to strike one at the hall' 
hour, the count- wheel or locking-plate is divided into 90 parts 
instead of 7 8, and all the nicks are made as long as the one 
o'clock nick in clocks that do not strike the half hour. 

Sometimes instead of a nick for every hour there is a pin 
with a slant side, which when the clock has struck the pro- 
per number raises the lever to catch the pin P in one of the 
higher wheels ; when warning is given the lever is raised by 
the going part a little higher to let the pin P clear the first 
stop S and stop at the second T, and when it is completely let 
off the lever drops and the count- wheel moves on and lets it 
fall low enough to clear both pins. This of course is exactly 


122 CLOCKS. 

the same in principle. The objection to the count-wheel 
method of striking is that if the clock is either caused to 
strike when it ought not^ or the striking part is allowed to 
run down^ or the clock put back past an hour^ it will strike 
wrong the next time^ and cannot be set right again without 
striking it round till it comes right ; and even this is not 
what every body knows how to do ; and it may be useful to 
those who have foreign clocks to be told that if you put 
your finger behind the clock or into the left side of it and 
lift up the only moveable thing you can fed, you can hardly 
fa3 to make it strike. There are several other modes of 
applying the locking-plate construction, which are occa* 
sionally used in turret clocks, but these are much the most 


90. The pl^ which has long been in use in all English 
house clocks allows the striking of any hour to be repeated 
as often as you please, or the clock to be stopped or put back 
past the hour,"^ or made silent, or to be run down, and 
still it will strike right as soon as it is made to strike again. 
Instead of a complete count-wheel or locking-plate there is 
only a portion of wheel L M, called the rack, turning on a 

* Borne docks will be found to offer a resistanoe to bdmg put 
back past the hour : if they do it cannot be helped ; but the lifting 
piece ought to have had its end twisted so that the pin on the reversed 
hour wheel.can pass the wrong way by pushing it aside. If a dock 
is put hack between warning and striking it will strike : this does not 
si^iify in an English dock ; but it will make any of the foreign 
docks, which all have the locking-plate movement, strike wrong, and 
therefore must not be done. 


centre 0, and having 13 or 14 ratchett shaped teeth on it. 
Another ndios ON of this fr^mentof avheelhas a pinN 
•et upon it; and on a wheel called the ttarieheel, which' 
will be described presently and tnnu in 12 hours, there is 
fixed a gradoated mail with 12 steps in it, such that when 
the honr hand points to 6 for instance the pin N can Call 
s^amst the 5th step in the snail to such a depth that the 
la^ can fiA through the space of S of the teetL Ihe 

124 CLOCKS. 

rack has a click D E, set on a stud at D, which will hold the 
teeth in any given place, a spring G pressing the rack to- 
wards the left hand. The arm D E of the click is prolonged 
to E, and there rests upon the lifting piece, which is the 
same as before, and has a stop T projecting backwards 
tlirough a wide hole in the clock frame, and so placed that 
when the lifting piece is down in its ordinary position, the 
pin P in the third wheel of the striking train can clear the 
stop T, but when it is raised it will stop the pin and con- 
sequently the train. When the clock gives warning, the 
pin R in the reversed hour wheel raises the lifting piece as 
before, which lifts the click, which lets the rack fall back 
as far as the snail allows it to go. 

Over the rack there is a sort of hook K, which is in 
fact a pinion with one tooth, set upon the projecting arbor 
of the second wheel, {Le. the one above the striking wheel), 
and at every revolution of that wheel, and therefore at 
every stroke of the hammer, it takes up the teeth of the 
rack one after another, and it is therefore called the 
gathering piece or pallet. When the clock is to strike, the 
lifting piece is let fall, which sets the pin P at liberty, and 
also lets the click fall on to the rack so as to be ready to 
catch the teeth and hold them as the gathering pallet 
gathers them up. After the pallet has taken up the last 
tooth, its tail, which is on the opposite side to the hook, 
falls upon a pin S at the end of the rack, and is thereby 
stopped from turning any farther, which of course stops 
the train. 

91. If the clock is to be capable of repeating the last 
"hour struck when required, the click FED is prolonged 


backwards and has a string put to it which comes through 
the clock case, and when it is pulled the clock will strike. 
But if the string is let go too quickly the click will catch 
the rack before it has fallen far enough, and the clock will 
strike too few ; and if it is not let go quickly enough the 
cUck will not catch the rack at the first stroke, and it will 
strike too many. Therefore the string ought to be put to 
the lifting piece, instead of the cKck, which can be done 
just as easily, though I have never seen it so done. If you 
find on pulling the string that the clock won^t strike, it is 
between warning and striking, and if you leave it alone it 
will strike of itself in a few minutes. 

92. The star wheel, of . which the construction is 
evident from the drawing, is turned ^^ih round at some 
time in every hour (it should be done before the work of 
the lifting begins, that the clock may not have too much 
to do at once) by means of a pin Q, on the hour wheel 
catching one of its rays. The pin does not indeed carry 
it through the whole distance it has to move, for by the 
time it has got about half way the ray at X in the drawing 
will have reached the angle of the thing called the Jumper, 
sliding up its left hand inclined plane, and as soon as the 
point of the ray passes the angle, the jumper, being pressed 
upon by a spring, will do the rest of the work and drive 
the ray still farther forward by means of the right hand 
inclined plane. The jumper also acts as a click to keep 
the star wheel steady; a common click would not do, 
because it would not allow you to put the clock back, 
whereas this one will allow the rays to pass up either of 
the inclined planes. In the commoner kind of clocks this 

1*^ CIX>OKS. 

star-wheel is dispensed with^ and the snail is set npon the 
twelve hour wheel (in which case it most be reversed) ; 
indeed the only use of the star-wheel is to relieve that large 
twelve hour wheel arbor or pipe of the additional weight 
of the snail; and if that pipe were reduced to a moderate 
me and the twelve hour wheel driveu directly by the train, 
as described in § 87^ the weight of the snail might be dis- 

93. Some clocks are so made that by turning a little 
hand which is placed above the dial the clock can either be 
made to remain silent, or be allowed to strike. This hand is 
so connected with the lifting piece that when it is turned to 
' silent' it pulls the lifting piece out of the way of the pin 
in the hour wheel by drawing its arbor forward, the pivots 
being made long enough for that purpose; and when 
turned to ' strike' it pushes the arbor back again. Some- 
times the striking is prevented by bringing a stop of some 
kind close up behind the pin S or some other pin set upon 
the rack, which prevents it from falling when the cUck is 
lifted, and so the gathering pallet cannot get past the stop 
S. This way is more simple than the other, but the other 
has this advantage that the clock can be made ordinarily 
silent, and yet it wiU strike whenever the above-mentioned 
string is pulled. 

I have seen a third way used, viz : the train stopped by 
a hook or pin .pushed into the way of the fly. This is a 
most blundering way of doing it ; for, as it does not prevent 
the rack from falling, the clock is almost certain to stop 
between 12 and 1 o'clock, from not having force enough 
to drive the deep step in the snail past the. pin N in the 

BEIili. 127 

rack tail, even if it is made with a bevelled edge for the 
purpose. In fact the clock in which I saw it had so 
stopped, and I was told it had frequently stopped before^ 
nobody knew why, I reKeved it from the dif&culiy in 
fature by taking away the ' silent' stop altogether. In a 
dock with the locking-plate movement however, this plan 
would answer perfectly, and is more simj^le than any other ; 
and probably the maker of this dock had borrowed the 
plan from an old locking plate clock, not recollecting the 

94. It is merdy a matter of taste whether a clock is 
made to strike on a bell, or on one of the spird sted 
springs of a very deep note that have lately come into use* 
Perhaps in a room the spring has a more agreeable sound; 
but they are not heard so far as a bell, and at a distance 
they have a thin sound, very unlike what they pretend to 
imitate, a heavy church bell ; and therefore for a clock to 
stand in a hall and strike for the whole house, a bdl is 
certaroly the best. 


95. House clocks are occasionally made to strike quar- 
ters. The machinery for that purpose is the same as for 
striking the hours, if we suppose two hammers to be put 
in the place of one, and an additiond set of pins on the 
striking wheel to raise them. If they are what are called 
dmg^dong quarters, on two bells, this wheel may be made 
exactly like the hour striking whed, only with two sets of 
pins set on its opposite sides so as to raise the two hammers 
dtematdy. But if there are four or more bells, the rim of 

128 CLOCKS. 

the striking wheel is spread out so as to form a cAime barrel, 
with pins sticking out of it like the barrel of a musical 
box, and these pins raise the hammers successively, which 
are all set on one axis. I may here remind clockmaters 
(as I have found it necessary to do so before) that though 
the repeating movement may be used for the quarters as 
well as the hour,, when the second, third, and fourth quar- 
ters are only the first quarter so many times repeated, yet 
when the tunes or changes are different the locking plate 
ought to be used ; for if the repeating plan is used and 
one quarter gets struck wrong or stopped, it throws the 
whole of the^tunes into confusion, the clock playing at tha 
next quarter the right number of changes, but the wrong- 
order, that is, a piece of the tune of one quarter, and a 
piece of another ; whereas if the locking plate is used the 
right tunes will always accompany the right number of 
changes, and any person will hear directly not only that 
they are wrong, but how they are wrong, and that they 
merely want striking by hand once or more to bring them 
right, which it is much more difficult for an unskilful per- 
son to ascertain when the tunes are wrong but the number 
right. The work for connecting the repeating quarter 
movement with the hour striking part is also more com- 
plicated than that of the locking plate, which is remarkably 
simple. But as quarters are so much more rarely put to 
house clocks than to turret clocks, I shall reserve the des- 
cription of the mode of connecting them with the other 
parts of the clock, as well as some other matters belonging 
to them, for the chapter on church clocks. 

96. Small spring clocks are occasionally made to strike 


the quarters only when they are wanted, as in the night, 
and to strike them after striking the preceding hour, just 
lis repeating watches do. In this case, the repeating part 
is not wound up periodically as in a common clock which 
strikes quarters ; but by pulling a string or pushing in the 
handle of the watch, you wind up a barrel containing a 
main-spring, which, when you let it go again, works the 
3triking movement in the usual way, according to the 
position given to the rack or locking-plate by the wheels of 
the going train. Bepeating watches are rarely, if ever, 
made in England, or at least by English workmen ; and 
from the quantity of work that has to be put in a small 
compass they seldom go very weU. Eor this purpose a 
small clock is much better; and such clocks are usually now 
made portable, that is with a balance instead of a pendulum, 
and go by the name of carnage clocks y and consequently 
they possess the advantage of being secure against the dis- 
turbances of housemaids. I shall have to mention them 
again in the next chapter. 


97. If you take the pendulum off a clock with a recoil 
escapement, you will hear it beat about as quickly as the bell 
is struck in an alarum ; and the striking there takes place in 
the same way as if a double hammer were put on the end of the 
crutch of a recoil escapement, just long enough for one end 
of the hammer to reach the inside of a bell at one extremity 
of the beat, and the other end at the other extremity. The 
manner of letting it off I shall describe presently; but 


130 CLOCKS. 

first it may be asked by aa observant person, how it 
happens that such a small weight as that of an alarum, fail- 
ing only a few inches> able to strike such a vast number 
of blows, when the striking part requires a' weight of 10 
or 12 lbs. to strike so few as it does. The reason is this : 
the common striking weight has to raise the hammer against 
a strong spring; and the force with which it strikes the 
bell depends upon the force of that spring, and the distance 
it is bent ; but in an alarum, the hammer is merely thrown 
from side to side by the action of the alarum weight, which 
has consequently nothing to overcome but the friction; and 
the velocity of the strikiug depends upon the proportion 
between the moment of inertia of the hammer and the 
momentum (or product of the weight and the fall) of the 
alarum weight. An attempt appears to have been made to 
produce a common striking part on this principle, but with- 
out success, because in that way it cannot strike with both 
force enough and slowness enough to be distinct. 

98. But an alarum is wanted to let off, not always at 
the same time, but at an arbitrary time. Por this purpose, 
all we have to do is to make the letting-off pin moveable on 
the wheel which carries it. If the common lettiug-off pin 
were made moveable upon the hour-wheel, we might make the 
clock strike when the long hand is at 50 m. instead of 60 m. 
And in like manner, if we put a letting-off pin for the alarum 
on a cap, or key^ which fits tightly on to the socket of the 
12 -hour wheel, we can make it let off with as much accuracy 
as is required at any portion of the twelve hours we please. 
This key, which carries the pin, has its socket prolonged 
forward, so as to carry a small dial or hour circle under the 



hour hand^ which always travels round with it; but the key 
does not fit so tightly that it cannot be moved by the hand 
so as to make a pointer in the hour hand point to any time 
we like on the alarum dial; and then it wiU let off at the 
time so indicated. There is no occasion for any stopping 
apparatus beyond a single lifting piece^ and a pin in the 
striking barrel^ which rests against the stop on the lifting 
piece when it is not raised high enough to let the alarum 
go. The string of the weight is attached to a wheel 
exactly like a recoil scape- wheel, either of the crown wheel 
or the plain kind, and when it is let off it runs till it is 
down. Of course it must not be wound up until within 
twelve hours of the time when it is intended to strike. 


99, This is mentioned incidentally in the parliamentary 
papers respecting the great clock for the new Palace, as 
being the invention of Mr. Whitehurst, of Derby. It is 
otherwise called the teU-tale clock, its object being to watch 
watchmen; whose tendency to sleep, like other men, in the 
night, had caused buildings protected by them to be con- 
sidered less secure than those which axe only protected by 
the ordinary precautions of mankind before they go to bed. 
In the watchman's clock there are pins sticking out round 
the dial, one for every quarter of an hour; and the duty of 
the watchman, in order to testify his own vigilance, is to go 
to the clock every quarter of an hour, and push in the 
proper pin, which only admits of being so treated for a few 
minutes : consequently, if in the morning any pin is found 

132 CLOCKS. 

sticking out, it indicates that the watchman was absent, in 
body or mind, at the corresponding time. There is a clock 
of this kind, on an improved plan, lately put up in one of 
the lobbies of the House of Commons. The face and pins 
are not open but inclosed behind a glass, and there is a 
handle at a convenient place outside the clock case, which 
communicates with a small lever standing over a part of the 
circle in which the pins move ; and as the pins are carried 
round in a sort of a moveable dial, it is evident that pulling 
the handle when any particular pin is under the lever, it 
may be easily made to push that pin in, and can do so at no 
other time. In fact, if the lever ended in a pencil, and the 
moveable dial carried a rim of paper set round the figures, 
puUing the handle would make a mark upon the paper, 
showing exactly the time at which it was done. But the 
paper would want renewing continually, and the pins do 
not, but are made to pass over an inclined plane at their 
back in another part of their journey (performed after 
the hour of inspection), and so to push themselves out 
again: for this purpose it is most convenient to make 
the revolving dial to contain 24 hours, so that there may be 
time enough in the morning to examine the pins before 
they are replaced by the inclined plane. 


100. If the pallets of a clock, instead of being worked 
by a pendulum, are attracted by pieces of soft iron, magnet- 
ized alternately by means of an electrical communication 
with the pallets of another clock, as they move to each side, 
or with the teeth of the scape-wheel successively, the two 


clocks will evidently go exactly together. Not only this, 
but if you take oflf the weight as well as the pendulum, and 
move^e pallets of a recoil escapement backwards and for- 
.they will drive the scape-wheel instead of letting 
3^/ escape, only they will move it, and consequently the 
hands, the wrong way; and, therefore, if the scape-wheel 
and pallets are reversed, they will drive the hands the right 
way. Instead of two magnets, one is sufficient, if the 
pallets are drawn one way by a small weight or a spring, 
and the other way by the magnet. This is the ordinary 
form of what are called electrical clocks. 

101. But still further, the whole train of the clock may 
be dispensed with, except the dial wheels, by making the com- 
munication take place at longer intervals, such as every half 
minute. The wheel of the minute hand would then require 
60 ratchett-shaped teeth, and a pair of large pallets to 
be put over it in the same way as the recoil pallets over 
a common scape- wheel ; and at each beat the wheel would 
move through the space of half a tooth. A click should in 
any case be added to secure the wheel from slipping back at 
the moment of escape. The advantage of a minute-hand 
moving a visible space at intervals, such as half a minute, 
will be pointed out under the head of 'turret-clock re- 
montoires,' § 172. 

102. Another way in which this may be done, is to let 
the dial work be driven by a weight or spring, wound up 
just like a common clock, and make the electricity let it off 
instead of driving it. This method is peculiarly adapted to 
clocks with large faces and heavy hands ; for in this way a 
dial, or any number of dials, of any size might be worked by 

184 CLOCKS. 

au escapement and train no larger than that of a common 
astronomical dock. The only objection to it is the difficulty 
of keeping a galvanic electrical machine in perfect action^ so 
that you can depend upon a beat never missing ; for it not 
oidy requires some attention to the galvanic trough^ but a 
small particle of dirt on the rim of the wheel, which is 
divided into conducting and non-conducting spaces, some- 
times interferes with the contact. Mr. Dent had an electri- 
cal dial, of the first construction just now described, going 
for some time in his shop, but he found it required constant 
attention to prevent it failing in this way. 

103. When, on the other hand, a large and strong clock 
is employed to drive a number of other clocks, magnetic elec- 
tricity may be used, the large clock by its own force pulling 
or rather sliding a strong permanent magnet out of contact 
with its armature at the proper intervals. Mr. Dent also 
kept a dial of this kind going for some time, making a small 
turret-clock movement work the magnet; and this experiment 
answered perfectly. Of course the clock which has to do this 
extra work should have some provision to prevent the inequa- 
lities of force produced by it from reaching the pendulum. 
It will be seen how this is to be done when we consider the 
subject of turret-clock remontoires. This general explanation 
of electrical clocks may be interesting to those who have read 
in the parliamentary papers on the subject, that it is pro- 
posed to have an electrical communication between the great 
clock at Westminster and the Observatory at Greenwich, 
and perhaps also to make it work the small clocks in vari- 
ous parts of the building. 

104. Now that the very convenient practice of setting 


all the clocks in the kingdom to Greenwich instead of local 
time is becoming, general^ it would be by no means an use- 
less employment of the electric telegraph to send the exact 
time at a certain hour every day or every week to all the 
telegraph stations on the various Unes of railway. The whole 
kingdom would thus have the advantage of an indication of 
the time as perfect as that afforded by the ball on the top of 
the Observatory at Greenwich, which is dropped every day 
exactly at one o'clock. Whereas at present, the time can 
only be obtained in places where there is not a good meri- 
dian instrument by the tradition of travellers or railway 
guards, which is equivalent to saying that it cannot be 
obtained at all with accuracy. 1 have sometimes found 
a difference of five minutes between the time of the public 
clocks in large towns less than 100 miles apart, and not ad- 
mitting of the excuse that it was the difference of local time, 
because, unfortunately, it has been occasionally the wrong 
way, i. e. the clock in the west has been faster than the 
clock in the east. 


105. These are merely common clocks, which let off a 
musical box, or a small organ driven by a weight, or a chime 
barrel in a church, in the same way as an ordinary striking 
part. A chime barrel is exactly the same as the barrel I 
before mentioned for playing a tune at the quarters, only 
much longer, having to play a longer tune and on at 
least eight bells, to each of which there are generally two 
hammers, because when a note has to be repeated quickly, 
one hammer could not fall and rise again quickly enough. 

136 CLOCKS, 

A musical box is the same in principle, only the music is in 
the springs raised by the pins on the chime barrel, instead 
of those springs being levers working hammers that strike 
on bells; and accordingly musical boxes are generally re- 
paired by watchmakers. Organs involve other considera- 
tions besides those of clockmaking, though many of the 
German clockmakers repair these clock organs; indeed 
they are hardly ever applied except to German clocks. 

Clocks are occasionally made to perform other feats, 
which I do not consider as properly belonging to the art of 
horology, but merely toys; and it is true of clocks, as 
of human beings, that if they profess to do a great many 
things, they seldom do them well. This remark, of course, 
does not apply to church clocks with chimes, which are on 
a large scale, and can have ample power and space to do 
the additional work. 


106. The theory of wheel-cutting does not belong more 
to clock-work than to any other machinery in which wheels 
are used, and it is too large a subject to enter into fully 
here. But it may be useful to make a few general remarks 
upon it, and I must refer the reader to other works, such as 
'Camus on the Teeth of Wheels,' ^Willis's Principles of 
Machinery,' &c., for the geometrical investigations by which 
these results are obtained. 

First, it is evident that the more teeth any two wheeli 
have which are to work together, the less friction there will 
be, and the less it will signify whether the teeth are acca- 


lately cut of the right shape. For if the teeth were infinite 

in number, and therefore infinitely small, there would be 

no sliding friction at all, and the wheels would merely roll 

upon each other; and on the other hand, the fewer the 

teeth of each wheel are, the farther they have to slide along 

the teeth of the other wheel. And as the same proportion 

must be kept between the wheels and pinions of a clock 

train, whatever the actual numbers are, the number of teeth 

or leaves in the pinions is at once an indication of the 

degree of labour which has been expended on that portion 

of the clock. In the best astronomical clocks no pinion 

has less than twelve leaves,** and sometimes as many as^ 

sixteen. In the striking part of clocks, there is no need of 

the same high numbers, because the uniformity, or amount 

of the force of the train, is of no great consequence (within 

reasonable limits) so long as it is always sufficient to raise 

the hammer. 

107. Another thing to be remarked in clock-work is, 
that the pinions are all driven by the wheels, except the 
pinion of the dial- work which drives the 12-hour wheel, 
and the pinion which drives the locking-plate when it is not 
placed upon the great wheel of the striking part ; whereas 
in machinery for raising heavy weights, the pinions gene- 
rally drive the wheels. And this is of consequence for the 
following reason. When two wheels act together, the teeth 
which are in contact may happen to be either entirely on 
one side of the line which joins the centres of the two 
wheels, or on the other, or on both sides. If the teeth have 
passed that line before they come into contact, the action is 
said to be after or behind the line of centres; ^ndivice vend. 


In this figwe, D,C, represent two 

te^Ii of the driving wheel, and d, e, 

two corresponding teethof the driveB 

wheeL Now, althoogh the; are here 

represented as inclined to each other 

at the same angle iu each ca«e, ;et 

any one mnst see at once that the 

friction between C sad e, after the 

line of centres, is very moch less than 

between D and d, before the line of 

centres : no degree of friction (short 

of adhesive friction) conld prevent G from driving e, if force 

enongh is applied to it ; whereas if the point of contact 

of D and if is a good way &om the line of centres, and the 

Bor&ce of D rough, a great force applied to the driving whed 

might merely drive the point of the tooth d into the side of 

the tooth D. The difference between drawing a walling* 

stick along the ground after yon and pushing it before you 

is an tllnstratiou of the difference between friction after the 

line of centres and before ; and a very moderate degree of 

roughness will make it impossible to pnsh it if it makes a 

large angle with the ground. It is evident therefore that 

machinery will go much more easily when the teeth of the 

wheeb are so arranged that all the wheels are driven behind 

the respective lines of centres. Now it may be shown 

geometrically, that with teeth of the common form a wheel 

cannot drive a pinion with leaves of sensible thickness witii- 

oat any friction before the line of centres, if the pinion has 

less than eleven leaves. This, therrfore, is another reason 

for using no pinions iu delicate clocks with leaa than twelve 


leaves, eleven being an inconvenient number, and of course 
the higher the number the better. 

108. There is, however, a kind (rf pinion^ in which 
(when driven, but just the opposite when driping, for which 
therefoie they are unfit) the action will always be after the 
line of centres, however few its leaves may be : it is a pinion 
in which the leaves, instead of being radial or of the leaf- 
shape, are round pins set in a ring round the axis; and 
from their appearance, these are called lantern or b(w jnmans. 
I have drawn the pinion of the scape- wheel in p. 79 as a 
lantern pinion. Of course, in these as in all other pinions, 
there will be less friction with many leaves than with few; 
but whatever friction there is will be entirely of the more 
harmless kind. The Dutch clocks always have these pinions, 
which are merely bits of wire stuck into wooden arbors, and 
«^ i. U» »«n wh, fl«, «U go «.h « ««o™.. of « 
that would completely stop a clock with the common 
pinions, and their pinions are the last thing in the dock 
to wear out. The Erench have also long used them in 
their turret-clocks, and it is stated in one of the parlia- 
mentary papers, by Mr. Dent, that they have been found to 
show no signs of wearing after sixty years: whereas in 
clocks made wholly with teeth, if not very accurately made, 
the pinions and the wheels that work them will be found to 
have worn themselves out of shape in a few years. I have 
seen a report made by a man who was sent to examine a 
turret-clock, made by a London maker, only fourteen years 
old, in which he stated that he found the pinions had 
been already twice shifted to get a new bearing for the 
teeth, and that they were now thoroughly worn out. Of 

i*0 CLOCKS, 

coarse this was an extreme case ; bat it would not have 
happened with lantern pinions and equally ill-cut teeth, 
because they would only have had the more moderate fric- 
tion of the action after the line of centres. 
• 109. Another great advantage of these pinions is, thai 
there is considerable risk of spoiling solid pinions, espe- 
cially large ones, in hardening them. I have seen a large 
pinion with half the leaves flown completely off in harden- 
ing ; and of course the risk of this adds to the expense of 
those that are sound, and also very often causes them not 
to be hardened at all, which is just one of those things for 
which you can have no security except the character of the 
maker. Whereas the lantern or box pinions never break, 
and are made merely by driving hard steel wire pins into 
the holes in the box, which are bored with great quickness 
and accuracy by a sliding drill and a dividing engine. Con- 
sequently, besides their other advantages, these pinions are 
probably, and when they are of large size, certainly, cheaper 
than solid steel pinions, which require cutting, hardening, 
and polishing, and often break when they are done. 

It must be remembered by those who are not used 
to them, that lantern pinions require the teeth that drive 
them to be cut of a different shape from teeth that drive 
common leaves, the teeth that drive the common leaves being 
epicycloids traced by a circle of >4a^*the radius of the pinion 
rolling upon the driving wheel ; while the teeth to drive 
pins are epicycloids traced by a circle of the same size as the 
pinion itself. No doubt very few teeth, especially in small 
clocks, are really epicycloidal, but some portion of circular 
arcs pretty nearly coinciding for a short distance with an 


epicycloid ; and where the teeth are numerous, and there- 
fore small, there is very little difference. But wheii the 
teeth are large, the difference is sensible, and there seems no 
reason why the cutters of the teeth of large wheels should 
not be made of the proper shape more frequently than they 
^e. It ought, moreover, to be remarked that, whether the 
teeth are circular or epicycloidal, the practice of topping, or 
turning off the tops of the teeth by way of correcting the 
depths is entirely wrong, because it takes off the most curved 
part of the teeth. If the teeth are cut right, and the depth 
is found to be wrong, the centres are wrongly placed, and 
should be altered, not the teeth, which are innocent. 

110. It is not indeed quite true that the action of a lantern 
pinion begins when the acting side, but when the centre of 
each pin, is on the line of centres ; and therefore the action 
is not so completely after the line of centres as a leaved 
pinion of twelve or more teeth. But practically this is no 
objection to their use ; first, because none but the most ex- 
pensive clocks have pinions of twelve, and it is not in clocks 
of that kind that there is much to be gained (as regards fric- 
tion) by using these pinions ; and secondly, the advantages 
of easy construction wouM still remaiQ on the side of the 
lantern pinions, especially when they are large. The scape- 
wheel of the Exchange clock is made in this way, although 
the pins are leaf-shaped, because it has twenty of them, and 
is therefore so large that it was found impossible to harden 
a solid pinion properly. For the same reason Mr. Vulliamy 
makes his large pinions of gun-metal, as hard as it can safely 
be used, which however is much softer than hard steel, or 
even cast iron. Thirdly, when you come to pinions of high 

142 CLOCKS. 

numbers, the angle corresponding to half the thickness of 
the pin is so small, that it is of no consequence whether the 
action begins that anall distance before the line of centres, 
or exactly at the line, for the nearer jou come to that line^ 
the less friction any deviation from it produces ; and there- 
fore making the pins of the Exchange scape- wheel of the 
leaf-shape instead of round was quite an unnecessary and 
very expensive refinement ; and Mr. Dent now makes all 
the pinions of the going trains of his turret-clocks lantern 
pinions, with round and hard steel wire pins, and the ease 
and smoothness of their motion is remarkable, even in 
a case where it was necessary to use a pinion as low as seven, 
which will be described in § 177. 



Chapter II. 

111. I have already said that the only mechanical dif- 
ference between a watch and a dock^ is that a watch will 
go in any position^ but a clock only in one ; for although 
timepieces in cases like small bracket-clocks^ only with 
a balance instead of a pendulum^ are popularly called 
docks^ they are really watches^ differing only from common 
watches in tiieir size, and in the balance being set horizon* 
tal, or in a plane perpendicular to the planes of the rest of 
the wheels. Such clocks have of course the advantage of 
being portable, and* are called carriage clocks. But imless 
they have compensated balances, they are mudi more affected 
by changes of temperature than pendulum clocks, or watches, 
which are kept at a pretty uniform temperature by being 
earned in the pocket, as will be explained in § 124. 

112. If you take a common spring dock, with its 
wheds and escapement arranged exactly like that of which 
a drawing was given at p. 49, and instead iA the crutch for 
the pendulum put a whed of the nature of a fly-whed, this 
will be v»y nearly a watch. Not quite, however; for it still 
differs from a watch in this, that the fly-whed ot balance has 
no spring, and so its vibrations wiU depend only upon its: 


own moment of inertia, and the. force of the train. And 
this last dependence is just what we want to get rid of, or 
to compensate in such a way that comparatively large vibra- 
tions of the balance shall be performed in the same time as 
smaller ones. Dr. Hooke ascertained, among his numerous 
other discoveries and inventions, that the vibrations of a 
spring in large or small arcs are very nearly isochronous. 
The reason of which is that the force with wliich a spring 
endeavours to resume its position generally varies as the 
angle through which it is bent, with the exception before 
mentioned in § 80 ; and a body acted upon by such a force 
as this (as a pendulum moving in a cycloid is) must oscil- 
late through any distance in the same time. Consequently 
if one end of a spiral spring is fixed to the frame in which 
the balance-arbor turns, and the other to the balance itself, 
in such a position that when the spring is at rest the pallets 
stand half way between each beat, and the balance is set in 
motion, it will go on vibrating in very nearly the same time, 
whether the force of the escapement drives it through a 
small arc or a large one. In fact, with pallets of this kind 
the watch will set itself going, because a tooth of the scape- 
wheel must always be pushing against one or other of the 
pallets, and the spring offering scarcely any resistance in the 
neutral position (unlike a heavy pendulum) the force of the 
escapement is suflBcient to move the balance through the 
angle of escape ; and having once begun, it wiU of course 
go on. Hence it is that with. most escapements a watch 
cannot be stopped, if it is clean and in good order ; for as 
the impulse is generally given at the neutral position of the 
spring (as it ought to be, for the same reason as in clocks). 


the force requisite to give tKe impulse is generally sufEicient 
to start the balance. The same thiug happens in a clock 
when yon take the pendulum off, and the esci^ement has 
only the slight weight of the crutch to move. 

lis. The spiral aprings used for escapements are of two 
kinds; the common one lying all in one plane, and the 
other, used for chronometers, in which size is not sacrificed 
to degance, is like a wire wrapped round a cylinder, as the 
string of a dock ia wrapped round the barrel ; since this is 
found to be the best form of the two, perhaps on account of 
every ring (so to speak) of the spiral having the same radius, 
and the sprii^ acting therefore more uniformly. 

114. The kind of watch that 1 have been describing is 
the old vertical watch, so called because the scape-wheel 
stands vertically when the other wheels are horizontal, or 
the watch is lad flat. This kind of escapement, as before 
mentioned, lite the common recoil escapement in clocks 
with anchor pallets, loses as the arc of vibration decreases. 


115. The mode of regu- 
lating this and all watches 
is by altering the fixed point 
of the spiral spring, so as 
to make the acting part of 
it shorter or longer, and 
therefore faster or slower in 
its vibrations. This is done 
as follows. A B D is the 


springs or rather a portion of it^ as I have stopped it at D 
to avoid confusion. The outer end of it is pinned into a 
cock set on the firame at A. The pointer C B E turns on 
an annular or hollow pivot at G, the hole being left in ltd 
oentre for the arbor of the balance^ which is called the verge, 
to go through. At B there are two small pios close to- 
gether^ between which the spring is passed^ and which 
therefore determine the point from which it begins to bend. 
Therefore as the pointer or regulator G B £ is moved towards 
the right in this drawings it makes the spring vibrate fester^ 
and to the left^ slower ; and this is done without affecting 
the neutral position of the spring. If the watch still goes 
too slow^ when the regulator is moved to the right as far as 
it can go, the spring has to be taken out and shortened by 
putting the outer end farther through A and drawing the 
inner end farther out through the other cock by which it is 
pinned to the balance ; for as a given angle corresponds to 
a smaller length at the inner end of the spiral than at the 
outer, the spring will on the whole be shortened by this 
operation, while the neutral position of it with regard to the 
balance remains the same. In order to alter the length of a 
cylindrical spiral spring, either its curvature must be altered 
a little or the point at which it is attached to the balance. 
But, in fact, chronometers are not regulated in this way, but 
by altering the in«rtia of the balance by screws with heavy 
heads, called timing screws. The vertical escapement ii 
only now used for the commonest watches, both on account 
of the thickness it requires^ and its inferiority to other 



116. The escapement which is far the most frequent 
in good English watches is the lever escapement, or, as it 
used to be called, the detached lever, to distinguish it from 
another form of it, now disused, called the rack lever. If 
you put a rack, or a few teeth of a wheel, of which the 
pallet arbor would be the centre, on the end of the crutch 
of a clock dead escapement, and let this rack work a pinion 
set on the verge of a balance, such as I have just now 
described, the vibration of the balance would cause the 
pallets to move -as they do under the influence of a peridti- 
lum. And this was the rack-lever movement. 

117. It was however liable to stop if the balance was 
accidentally stopped at the neutral position, on account of 
the friction between the rack and the pinion ; and moreover 
the lever or crutch and pallets were carried farther the far- 
ther the balance vibrated. This as we have seen cannot be 
helped with a pendulum, on account of the small angle 
which it moves through ; but it can with a balance, since 
that moves through such a large angle that the arc des- 
cribed by a pin set a little distance from the verge will cut 
the arc described by the end the lever so as to include a 
very sensible depth between them. Consequently if the 
end of the lever merely has a nick in it, and the verge in- 
stead of beiDg a complete pinion has one tooth or pin that 
will fit into the nick, the lever and pin will act together as 
a wheel and pinion for a short distance in the middle of the 
vibration, but as soon as the pin has got out of the nick 
the balance may turn as far as it pleases without moving or 

H 2 


being even in contact with the lever; and when it returns the 
pin wiU go into the nick again and first move the dead part 
of ihe pallet off the tooth of the scape- wheel, and ^sa receive 
the impulse and leave the opposite pallet with a tooth test- 
ing upon its dead part. Therefore this waa called the de- 
tacked lever, as the lever is detached from the balance except 
during the middle of the vibration. The pin is usually 
made of a jewelj and it works with so little friction that 
Tou can hardly (if at all) stop the watch, vxj more than a 
clock with a dead escapement when the pendulum is taken 
i)H'. The pallets are also always made of jewels in good 
watches. In 
this drawing 
the balance is 
put beyond 
its real dis^ 
tance from the 
order to show 
theother parts 
more plainly: 
it is exhibited 
at the middle of the impulse. The practical advanti^s of 
tills movement are that it is not only avery good one, being 
like the dead escapement of a clock, uid without the dead 
friction, which is very nearly removed by the detachment of 
the balance from the pallets, but it is moreover easy to make 
iind safe t« wear ; and if the watch gets such a Ml as to 
break the ve^, which is always the first thing to break, 
it can be mended for a few shillings, the verge in this 


escapemeat being Dotbing more tban a plain arbor^ wbich 
camefl the pin set upon a small collar. Whereas in the 
two nest escapeHwnts the vei^ itself is comparatiTely com- 
plicated, and expensive to make and mend. 


118. It is curioos that this, which was invented by an 
English clockmaker, Oraham, is the escapement pnt to nearly 
every foreign watch, whereas the rack lever, which was the 
foundatioii of the detached lever, the principal English es- 
capement, appears to have been invented in France, though 
the detached lever was invented by Mudge, a famons chro- 
nometer maker here. There. is nothing in clocks correspond- 
ing to this escapement, and it will be best understood at 
once &om the drawing. The vei^ is spread out about ils 
middle into a portion of 
a hollow cyhnder, of 
which the section is 
sliown at Afi. The 
scape-wheel teeth are of 
the shape of a ' flat iron' 
set upon a pin, as shown 
in the second flgore, and 
just long enough to be 
contained with the cy- 
linder. A tooth is here 
represented as just es- 
caping iiom the inside of tue cjlmdet and giving the iiu- 


pulse ; for as the balance turns to the right"^ the sloped part 
of the cylinder at A comes up to the point of the tooth, 
and so the tooth in making its way out, gives an impulse to 
the balance ; and then the next tooth strikes against the 
outside circular part at B, and rests on it until the balance 
turns the other way, and then that tooth makes its way into 
the cylinder giving an impulse in so doing; This escapement 
therefore is not detached, as a tooth is always pressing 
either upon the inside or the outeide of tiie cylinder, and as 
in the common dead escapement, a tooth is always acting on 
the inside of one pallet or the outside of the other. This 
is not unlike the pin wheel dead escapement in this respect^ 
that it is not very material that the teeth of the scape-wheel 
should be exactly at the same distance from each other^ 
since they act successively. But is very unlike it in ano- 
ther respect, that it is very expensive to mend if broken. 

One curious fact in this escapement is worth mention- 
ing : it is found that the sharp teeth of the scape-whed 
cut the cylinder less if they are made of steel than of 
brass. I have heard of no satisfactory way of accounting 
for this ; and it seems that there is yet room for a good 
deal of discovery in the theory of the wear of two metals 
working together, in which several of the known facts are 
different from what were anticipated, especially the necessity 
for soft bearings, made only of block tin, which is nearly as 
soft as lead, for railway axles orjoumah ; while on the other 
hand I understand that brass, aind even gun-metal, is aban- 

* Perhaps it may be as well tO'explain, that a wheel is said to 
turn to the right, when, if you took hold of it with your right hand 
in its natural position, you would turn your thumb to the right. 


doned for tlie bearings of lathe sprndlea in fevoor of cast- 
iron or steel. There is no doabt, however, that iron and 
Bted working together reqtiire freah oil more frequently than 
brass and steel; and brass scape-wheels are nearly always 
used for the dead escapement in clocks, and the lever 
escapement and all others except the horizontal in watches, 
which is just the one in which we should have expected the 
points of brass teeth to be worn out sooner than in any 
othex, instead of wearing the hard cylind^. 

119. The daplex escapement is so called because the 
scape-wheel has a doable set of teeth : viz., a set of long 
teeth, which merely rest gainst 
the verge of the balance except 
when they are escaping, and a 
8C(t of short teeth or pins to 
give the impulse at the time of 
escape. The long teeth are so 
made that their forward edge is 
a radius of the scape-wheel and 
a tangent to the small circle 

with a notch in it j consequently 

that circle or cylinder, which is a portion of the vei^e, can 
torn to the right, with no other effect upon the long teeth, 
than that the friction against th^ is rather greater as the 
opening of the notch passes, than when they are merely 
pressing upon the cylinder ; but as the cylinder is very 
■small, and the teeth long, and the notch narrow, this extra 
friction is very little. When the balance returns, the point 
of the tooth Ms into the notch, and goes along with it, and 


SO escapes^ and at the same time one of the pins catches the 
long tooth projecting from the verge^ and gives the impulse. 
It will be seen that the principle of this escapement »» 
not unlike that modification of the dead escapement for 
clocks^ which I suggested (45)^ and which if it can be made 
accuratdj enough with moderate care^ ought to answer 
equally well ; for here the dead friction is that of long teeth 
acting upon a cylinder of small radius^ while the impulse is 
given by short teeth acting against a long radius. !For this 
reason^ and also from the action being so direct that it 
does not require oil^ the duplex escapement is a very good 
one; but^ as may be inferred from this description of it^ it 
requires unusual care in making and putting together, and 
also in carrying about ; for if, by an accidental twist of the 
watch^ the balance should h^^pen not to vibrate tax enough 
to let the notch get past the long toothy the escape cannot 
take place^ and the balance loses its impulse for that vibra- 
tion ; and as it only receives an impulse at every alternate 
vibration, unlike the three preceding escapements, it will 
probably not recover itself and will stop. It is therefore 
unfit for ordinary wearersof watches, and above all for those 
who in winding up their watch turn it nearly as much as^ 
they do the key : a very mischievous practice with any good 
watch, as it deranges the arc of vibration for the whole time of 
winding up, and some time afterwards. It is not perhaps very 
easy to avoid it altogether with an ordinary key; and there- 
fore I should recommend persons who have really good 
watches to get a key made with a straight wooden handle 
like a thickish pencil. You can then wind up your" 
watch, keeping it quite steady, by a few twirls of this 
pencil between your finger and thumb, and in a quarter of 


tix time that a common key takes. Tb.e watchmakera them- 
selvea always use these pencil keys. 

120. The perfection of all the watch eacapemeiito is 
tliat which has from its use acquired the name of the 
chronameter movement, but was originally called the de- 
tached, the balance being entirely detached from everything 
else except just at the time of action : in other words there 
is no dead Miction; and moreover the impulse is given 
direcUy, and nearly at right angles to the line of centres of 
the balance and scape wheel, as in the duplex escape- 
ment, instead of obliquely as in the vertical, lever, and 
horizontid esca^meuts ; and therefore it is not subject to de- 
rtmgemesit fay Uie variable stale of the oil, as it requires none. 
The teeth of 
the scape- 
whed, in- 
stead of rest- 
ii^ against 
a verg^ rest 
against a 
stop S set 
upon a lever 
a b. This 
lever having 
only to move 

through a very snmll angle is set upon a small and stiihsli 
spring at a instead of a pivotj which allows it to move just 
as a pendulum spring doea a pendulum. The other end of 


the lever has a weaker spring h also screwed to its inr 
side .(that is the side nearest the scape wheel) and pro- 
jecting a little beyond its end. On the verge there is a 
small tooth or cog c which can pass the spring in the 
direction towards the scape wheel, by merely pushing it 
aside as shown in the enlarged drawing of that part, as it 
has room to bend in that direction; but in going the other 
way the spring cannot bend, and therefore the tooth c 
carries the lever with it, moving on its own spring a; and 
in so doing the stop S is pushed out of the way of the tooth 
of the scape wheel, which is therefore let go ; and at that 
moment the long tooth or snail end attached to the verge 
(which corresponds exactly to that in the duplex escape- 
ment), comes into such a position that the tooth of the 
scape wheel catches it and gives the impulse. 

It is evident that this also is an eseapement that re- 
quires considerable care in making and using. However 
it is never applied to any but chronometers, or the very- 
best watches, and those of considerable size, and persons 
who possess them are aware that they are instruments of a 
very different class from an ordinary watch. It is men- 
tioned in the Encyclopsedia Britannica that a maker of 
the name of Owen Eobinson made use of long teeth for the 
stopping part, and short ones for the impulse, as in the 
duplex escapement, but the contrivance has not been gene- 
rally adopted. Probably the friction of unlocking the 
common teeth is so small that the two sets of teeth were 
not worth the additional trouble, and the long teeth in- 
creased the inertia of the balance. 

121. It may be observed that as the lever is not conn- 


terpoised^ tbe work to be donje in unlocking is different^ 
according as the watch is placed vertically with the weight 
of the lever pressing towards the scape wheel, and therefore 
having to be lifted at unlocking, or the other way up, when 
the weight of the lever helps to unlock it, or horizontally, 
when the weight acts neither way. Consequently it is desira- 
ble to keep a chronometer always in one position ; and that 
position is the horizontal one for another reason, which 
applies to several things besides chronometer balances. 

If you take a small portable clock with a balance 
heavier than any watch balance, and lay it on its back ot 
side, and observe the vibrations of the balance, you will see 
that they are much less than when the clock stands up, so 
that the balance is horizontal. The reason is that a pivot 
standing upright and with its point resting on a hard 
surface, and merely kept in its place by the hole in which 
it is placed both at the top and bottom, moves with much 
less friction than the same pivot set horizontally in circular 
pivot holes. Por, however thin the pivot may be made, 
of course its point can always be made smaller, and in fact, 
if the wheel is Ught, it may approach indefinitely near to a 
point, in which case there would be no friction, at least none 
which the weight of the pivot and its wheel would augment, 
the lateral friction against the sides of the hole being inde- 
pendent of the weight resting on the vertical pivot. Ship 
chronometers are accordingly kept horizontal by being hung 
in gmbala (see. next page), which are in fact an universal 
joint, the chronometer having two pivots E, W, which move 
in holes in a large ring having other pivots N, S, at right 
aligles to E, W, which turn in holes in the sides of the box. 


The fly of the stziting 
part of very large clocks 
is in like maimer some- 
times placed with its axis 
vertical, being connected 
witli the train by bevelled 
wheels, partly in order to 
avoid the friction on the 
pivots of such a heavy fly 
movingwith great velocity, 
and partly to save room. 

122. Another remarkable application has latdy been 
made of this principle of nsing vertical pivots ; viz. in the - 
ship compass. Everybody knows the common way of doing 
it, viz., fixing the card and needle upon a conical cap, or in- 
verted cup, standing on the top of a sharp spike. Thia, <d 
course, has very little friction, when merely revolving level; 
but then it is liable at sea to great disturbances which some- 
times render it useless ; and moreover these vertical oscilla- 
tions soon wear the point blunt. Mr. Dent has therefore 
lately applied the chronometer suspension to compasses; 
that is, the axis goes right through the card like the axis or 
ve^ of a watch balance, and rests at the bottom on a 
jewel or hard piece of steel put under the pivot hole, and 
at the top in an ordinary pivot hole set in a frame above 
the card, the whole apparatus, as usual, being swung in 
gimbals in the same way as the chronometer just now des- 
cribed. There appear also to be several collateral advantages 
in this kind of suspension besides its steadiness ; such as 
the power of making the compass as tiuggUh as yon please 


by means of an adjustable spring pressing against the side 
of the upper pivot, and the power of inverting the compass 
to ascertain the error of colUmatian. , However, it wonld be 
out of place here to dwell at any length on the compass 
suspension, which is only introduced as an illustration of the 
principle of suspension of the marine chronometer. 

I said just now that the lower pivot of this compass 
rests on a hard piece of steel or a jewel. In the escapement 
of all good watches, and in the best clocks, there are what 
are called end-steps; that is, the pivots are not kept in 
their place endways by their shoulders, but by stops, of 
metal in clocks, and jewels in watches, against which 
the pointed ^ids of the pivots rest, not of course tightly, 
but suffi(»enfly dose to allow only the necessary shake 
or freedom. The necessity of this freedom, both end- 
ways, and sideways, of all the pivots in clock and 
watch work, is one of the points in which it differs from 
common engineering; for in other machines there is 
generally force enough to spare, so that a slight degree of 
tightness in a pivot does not signify, especially if plenty of 
oil is used. But in the going part of a clock such an 
occurrence would probably stop it. Besides the end-stops, 
the pivot holes for the balance, and the scape-wheel, are 
made in jewels in good watches ; and beyond these there 
appears to be no use in jewelled holes ; and watches that 
are called jewelled in eight or ten holes are often inferior to 
those which are only jewelled in the four I have mentioned. 

123. It requires some experience and quickness to note 


the exact second at which any phenomenon takes place 
which you have to employ your eyes upon. Astronomers 
do it by looking at tl\eir clock at some given second^ and 
then counting the beats from it by ear as they look through 
the telescope ; and experienced persons can do this to a 
t^ith of a second. But for the purpose of making this 
more easy to ordinary people, watches have been lately 
made with a large second-hand, carrying some colouring 
fluid on its end; and there is a small pia which you 
press with your finger at the. moment when you observe the 
phenomenon, and this presses the point of the second-haad 
upon the dial, and so makes a mark, which you can after- 
wards examine at your leisure, and without the necessity of 
employing your attention on counting the seconds. The 
same thing used to be done by the pin stopping the watch 
when it was pressed, which of course made it wrong after- 
wards. In a stop watch, of either kind, you can only 
denote the time to that fraction of a second, which the 
scape-wheel happens to beat; and therefore a watch with a 
lever escapement will be better for this purpose than, one 
with a detached or duplex movement, since in them the 
second-hand only moves at every alternate vibration of the 

In other respects watches differ so Uttle from clocks in 
the principles of their construction, that it is unnecessary to 
say any farther on that subject. But there stiU remains to be 
noticed a condition which is even more essential to the 
accuracy of their performance than to that of clocks. I 
mean the 



124. The balance of a watch requires compensation 
much more than a pendulum^ if it is to be exposed to 
such changes of temperature as chronometers are^ though 
watches carried in the pocket are not subject to much vari- 
ation of that kind^ unless they are left exposed in cold 
nights. The variation in the elasticity of the spring, which 
affects in a small degree the vibration of a heavy pendulum 
whose time is mainly determined by gravity (65), affects a 
light balance whose time of vibration is mainly determined 
by the force of the spring a great deal more. I extract 
from a smaU pamphlet on compensation balances, published 
by Mr. Dent, the following results obtained from a glass 
disk as a balance, which was used for the experiment 
on account of the moment of inertia of such a disk being 
less affected by temperature than a metallic one. 

Temp^atL. Yibrations in au hour. 

32« 3605-7 

66*^ 3598-2 

100^^ . . . . - 3589-7 
(The proper number being 3600). Supposing, therefore, 
that the balance had been adjusted to go right at 32^, 
it would have lost 7*54 and 8-5 seconds respectively for 
the first and second increase of 34^, which is equal on the 
average to more than 3 minutes in the day; whereas we saw 
(57) that a common iron wire pendulum would only lose 10 
seconds in a day, with about 1^ more for the spring, under 
such an increase of heat of 34^. And if a metal balance 
had been used instead of a glass one, the loss would have 


been still more^ on account of its expansion. I believe it is 
not far from the truth to say that the effect of the expan- 
sion of the metal in a balance is as much less than that of 
the weakening of the springs as the effect of the weakening 
of the spring is less than that of the expansion of the sted 
rod in a 1 sec. pendulum. Any body may easily satisfy himself 
that the variation in the elasticity of a watch-spring vastly 
exceeds in effect that of the spring of even a short pendu- 
lum^ by putting a common watch and a bracket*clock with a 
pendulum together in a cold room in winter, and adjusting 
them (if necessary) till they go pretty well together ; then 
shut them up in a hot room for a day, and you will see that 
the watch has been left; several minutes behind by the dock, 
in consequence of the greater effect of the heat upon the 
balance spring than upon the pendulum spring ; for as far 
as the expansion of the metals is concerned, the balance and 
the pendulum rod will expand about equally. 

125. The difference in the amount of compensation 
required by a balance and a pendulum is so great, that it 
cannot be effected exactly in the same way, though all the 
methods that have been invented depend upon the same 
principle, of making small weights attached to the balance 
approach nearer to the centre as the heat increases, so as to 
diminish the moment of inertia, or the resistance to the 
force of the spring. 

It is true of a balance, as of a pendulum, that if 
it were composed of rods or spokes without weight and 
a rim consisting of a single heavy line, or even a line with 
weight in only one point, the tiine of its revolution would 
depend (the force being the same) solely on the square of 


the radios of the rim. Bat aa the mere expansion of the 
wheel itaelf prodnces a small effect compared with the 
decrease of force in the spring, it is evident that we most 
have some more violent method than the ttraight fincaTd 
expansion of one metal over that of another, which will do 
in a pendulum, to produce the requisite effect in a balaoce. 
And this effect is produced accordingly by the curvilinear 
expansion of one metal upon another. 

If you fasten together tightly, as by soldering, a bar 
of iron and of brass, and heat them, the brass expanding 
more than the iron will evidently bend the iron inwards. 
The way, therefore iu which a compensation balance 
is made is this. A ring of steel is made with a bar 
across the middle; outside the ring is cast another 
ring of brass in such a way as firmly to adhere to 
tile steel, in fact to be brazed to it; this ring is then 
filed up to the proper size, and then a broadisb cut is 
made through both rings at each 
of the comers A, B ; and finally a 
few screws with heavy heads are 
set in various places near the end 
(rf each portion of the cut ring. 
Consequently, as the heat dimi- 
nishes the force of the spring, it 
expands the outside brass curves 
more &an the inside steel ones, and so their ends, with the 
additional weight of the screws, bend inwards towards the 
centre of the balance, and its moment of inertia is dimi- 
nished. The proper adjustment of these screws is, as may 
be supposed, a very delicate operation, and requires a great 


many trials at differeut temperatures^ and can therefore only 
be done in winter^ or by means of freezing mixtures ; and 
is consequently very expensive. And many watches with 
compensation balances are sold^ which have never been ad- 
justed at all, and nobody can tell by merely looking at 
them whether they have or not. They may be somewhat 
better than completely uncompensated balances, as the 
screws have probably been put in according to the common 
rules for a first approximation; but this is a matter in 
which no security can be had except the reputation of 
the vendor, unless he will allow the watch to be taken away 
and tried in different temperatures: any person, of ordi- 
nary observation, who has the means of resorting to a 
good clock can easily try whether the compensation is 
tolerably complete, or is over-done, as it sometimes ia. 

A similar contrivance has been applied to compensate a 
pendulum : two compound rods of this kind, with the brass 
«ide downwards, being made to project from the bob, and 
carry balls at their ends, which therefore rise as the bob falls 
and the brass expands. They have, however, never come 
into general use, aad all such projections from a pendulum 
are objectionable on account of their tendency to twist if not 
placed exactly in the plane of vibration; and it would be 
unsafe to trust to a compensation of this sort without actual 
trial, which would make it at least as expensive as a zinc 
compensation; and on a large scale the compound bars 
would probably split themselves asunder instead of bending. 

126. Still however the compensation is not completely 


e£Eected in this way. It has been fonnd that if it is ad- 
justed for two extreme temperatures^ such as S%^ and 100°, 
the watch will gain at mean temperatures, and if adjusted 
for mean it wiU lose at extreme temperatures. The reason 
is this. The force of the spring is ascertained to vary as the 
temperature ; but the isochronism at different temperatures 
depends upon the ratio of the force of the spring to the 
inertia of the balance remaining the same, just as the time 
of a pendulum depends upon the ratio of the force of 
gravity to the radius of oscillation. Now by the ordinary 
method of compensation the weights approach the centre at 
a rate nearly proportionate to that of the increase of tem- 
perature, in fact rather more rapidly at low temperatures 
than high ones, which is just the reverse of what is wanted. 
Por the inertia of the balance depends on the s^are of the 
distance of the weights from the centre; and therefore in 
^er that it may always bear the same proportion to the 
force of the spring, the distance of the weights ought to 
vary more rapidly when they are near the centre than when 
they are far from it, that is more rapidly at high tempera- 
tures than at low ones. 

-To persons accustomed to mathematical formulae the 
nature of this result will be clearly exhibited by observing 
that if r be the distance of the weights M (which for con- 
venience we may suppose to be the whole mass of the 
balance) from the centre, the new moment of inertia, for 
an increase of temperature which causes an increase of r 
which we may call dr, will be Mf r'^- 2 ro[r+ d/j ; or the 

ratio of the new inertia to the old will be 1 + —^ — f- ( T/ • 



The reader may remember tliat in the case of com- 
pensated pendulums we neglected the quantity corres^ 

jtonding to (— ) because of its smallness ; but the com- 

Iiensation required for a balance spring is so much greater 
than that required for the expansion of a pendulum rod^ 

that the effect of the term ( ^ j now becomes sensible ; 

for^ as was stated before^ the object of these compound bars 
of balances is to make r vary much more rapidly than it 
could do if the metals acted only by their direct expan- 
sion. And it is this term yy) which renders necessary what 

is called the secondary compensation, which is of course a 
great deal smaller than the primary one which is represented 

by ^-ZL , but stiU is large enough to require a special contri- 
vance in very accurate dironometers. 

127. Or to adopt the illustration given in the pamph- 
let I have referred to, the variation in the force of the 
spring may be represented by a straight line ABC in- 
clined to the straight line of temperature at an angle re- 
presenting the ratio at which the primary compensation 

must advance; where- 
as the variation in the 
inertia can only be 
represented by a curve 
such as A £ C, which 
can be made to coin- 
cide with the straight 
Une at any two points 


ve please, but at no others. This figure will show that if 
we make them coincide at extreme temperatures the curve 
fells below the straight line at the intennediate t«mperstures, 
or the inertia is too little for the force, and the watch will 
gain ; whereas if we adjust it for any two adjacent tem- 
peratures it will lose for all above or below them. 

128. Several contrivances have been adopted to effect 
this object of making the weights move faster when they 
are nearer the centre. It will be sufficient to describe one 
of them in snch a way as to show the principle of their 
construction, which is, if compound bars are used, to give 
them such a curvature as will oarrj the weights at the ends 
in a direction more nearly coinciding with a radius of the 
wheel when they are nearer the centre than when farther 
from it. It will be easily seen from tliis drawing that if 
the curved pieces are made of 

brass and steel, with the brass 
inside the curves, they will 
carry the weights inward as 
the temperature increases, and 
that the motion will be more 
in a radial direction when they 
are comparatively near the 
centre than when far from it. 

129. The method just now described is Mr. Dent's- 
It appears from a report of the astronomer royal to the 
Admiralty, in a parliamentary paper lately published, that 
the same thing was done to a certain extent and the 
principle of it suggested by a Mr. Eiffe ; and that it has 
lately been done with great success by Mr. Loseby (several 


of whose chronometers have been accordingly purchased 
for the Admiralty) by means of mercurial tubes^ curved so 
that the expansion carries the mercury outwards, when its 
extremity is near the centre, more rapidly than it does when 
the mercury has reached a part of the tube more distant 
and more curved towards the centre. I have no means of 
giving a correct drawing of Mr. Loseb/s compensation 
tubes, but the principle of their operation is sufficiently 



Chaptbe in. 


The principal diflFerence between church or turret clocks 
and house clocks is in their size. But this difference makes 
it necessary to attend to some things in their construction 
which hardly require consideration in small clocks. More- 
over they cannot always be made in the same way, because 
they have to adapt themselves to various situations, and 
have very different amounts of work to perform, and are 
required to satisfy very different conditions as to accuracy 
and price. 


130, In all cases, however, the first thing to attend to 
is the steadiness of the suspension of the pendulum ; which 
may vary in weight from 1 to 3 or 4 cwt., according to the 
size and number of dials that the clock has to work. No 
rule can be laid down upon that point ; but for the reasons 
stated in § 39 the pendulum ought always to be as 
long and heavy as it convenientiy can be, and the greater 
the amount of friction, and therefore variation in force, the 
train is Uable to, from large dials, the greater necessity there 
is for a powerful pendulum. And the heavier and longer 



the pendulum i&, the more it will require a firm support ; and 
this is generally best obtained by fixing the cock which 
carries it to the wall behind the clock. The cock should 
be cast on a large iron plate^ which is fixed by bolts leaded 
into the wall. The bolts should have shoulcfers to receive 
the plate^ or else screwing the nuts on will tend to draw 
them out of the wall, And the bolt holes in the plate 
should be large enough to allow some motion for adjust- 
ment. For cheap clocks the cock may be merely a strong 
piece of cast iron^ leaded into the wall^ with a proper 
opening in it wide enough to carry the spring and cAope 
in the manner described for astronomical clocks (26)^ which 
is the only way of fixing the top of the spring firmly. 
But care should be taken to put the cock in so that the 
pin through the chops will lie quite level and also parallel 
to the wall, or at right angles to the pallet arbor, and only 
so much above it that the bend of the spring will come just 
opposite the end of the arbor. 

131. The spring should be broad and thin, not 
narrow and stiff, and not more than 4 or 5 inches long, 
from the chops to the pendulum top.* Sometimes, to 
prevent the pendulum from twisting, two narrow springs 
are used, separated by about an inch. But a single one, 
broader than the two together is better, not only because 
it ^tU affect the vibrations less by its own elasticity, but 
because, if the two springs are not exactly of the same length 

* Mr. Dent has found he can get from the saw-makers better 
pieces of steel for turret-clock springs than from professed spring- 
makers, and at half the cost. I have seen some specimens of each, 
and there is no question which is the best. 


and atrengthj they will tend to twiat the pendulum at every 
vibration. The chops may be as well of cast irou as of 
brass, and they should be screwed together through the 
spring as near the bottom as can be. In this drawing I 

have shown the chops and spring and a side view ol the 
cock and the plate belonging to itj and the lewises of the 
bolts. I have also given a front view of the fonn of plato 
which appears to me the best for obtaining steadiness with 
four bolts. 

132. Of coQise a compensated pendulum is better than 
an uncompensated one; but the expense of compensating 
a 14 -feet pendulum is an obstacle to their use. The Ex- 
change clock is the only one in England, if not in the 
world, with a compensated pendulum of that length. Mr. 
Dent, however, has lately begun to use 8 feet pendulums 
compensated, with bobs of nearly % cwt; and it appears 
that such pendultmis do not cost above £7 or £8 more thau 


a wooden one equally well made. It may be as well to 
remark that cast tubes of zinc are not found sufficiently 
solid to expand and contract as they ought to do; and 
consequently it is necessary to make the compensation 
tubes of se\F^al tabes oi the common sheet zinc soldered 
together, in order to obtain sufficient strength to carry a 
heavy bob without bending. 

133. If the pendulum is made of wood, whether deal or 
mahogany, it should be as straight in the grain as possible^ 
and free from knots, and not above half an inch thick and 
two inches wide (which is strong enough to bear several 
tons); for the thinner it is, the less the weight of the rod, 
and therefore the radius of oscillation of the pendulum, can 
be effected by its absorbing moisture, and the straighter the 
bob will keep it if it has any tendency to bend. It should 
be dry and well varnished several times over, and the ends 
ought to be well saturated and aH the screw holes, to keep 
out the damp. The wood should reach as near both to the 
top and the bottom of the pendulum as the necessary metal 
terminations will allow ; and with that view, perhaps, it 
is better to make the crutch with a pin to go into the pen- 
dulum rod than with a fork : the pin should be adjustable 
for beat in the end of the crutch. The pendulum should 
end in a point, and should have a degree plate under it 
marked to about 10'. It may be convenient to state that the 
length for a degree on the plate is very nearly -^th of the 
whole length of the pendulum, and no very great accuracy 
is necessary, as you only want to see how much the arc 

When an eight feet pendulum is used, it is generally 




better to elevate the clock, so that the pendulum is wholly 
within the chamber, than to let it hang down in a box 
below ; for otherwise the bob cannot in most cases be easily 
got at to regulate the pendulum, or to observe the arc it is 
swiliging, which affords an indication whether the clock 
wants cleaning or not. There can be a moveable pair of 
eteps or platform to wind up by, which when not used will 
go under the clock, and so take up no room, as a fixed plat- 
form would. I may add, that the screw for raising the bob 
ought always to be at the bottom of the rod, and not at the 
top, or, still worse, at the side of the bob. In a compen- 
sated pendulum it would, as a matter of course, be put 
at the bottom ; and in a wooden one it caonot be put at the 
top without making part of the rod of metal ; and if it is 
put, as it sometimes is, at the side of the rod, a little way 
above the bob, altering the screw has a tendency to bend 
the rod, so that the bob does not hang vertically upon it. 
Care must be taken in making compensated pendulums, 
both that the bob and compensation tubes do not turn with 
the nut, and also that when you hold the bob steady, turn- 
ing the screw does not turn the rod, and so twist the spring; 
and therefore the zinc tube must not rest on the nut, but on 
a collar, which slides on a square or some other uncircular 
portion of the rod, and the collar and the nut should present 
convex surfaces to each other, to allow the nut to turn more 
easily. Bespecting the shape of the bob, and the method of 
adjusting, I have already spoken in § 70, 71. 

134. There is a practice, not uncommon with country 
clockmakers, of hanging the pendulum to the wall at the 

side of the clock instead of behind it, and connecting it 

I 8 


with the crutch (which sometimes goes np instead of down) 
by a horizontal rod two or three feet long. This^ though a 
convenient arrangement in appearance, is a very bad one 
for the performance of the clock ; for it is evident that the 
friction, and the shake, of the two additional pivots must 
tend to diminish the vibrations of the pendulum, or in other 
words must require a greater force on the pallets to maLn- 
tain the proper vibration; and therefore, by virtue of what 
is stated in § 36, must make the dock go worse than if the 
pendulum were suspended in the common way, just oppo- 
site the pallet arbor. No doubt there may be cases in 
,M.Ut ^ h^, po^bl. ..^.^.lo»g'pe.M^i. 
any other way; and when it is necessary so to suspend it, 
the horizontal connecting rod should be as light as possible, 
and its pivots small and accurately fitted to the crutch and 
the pendulum, so as to diminish the friction and the shake 
in the pivots as much as possible. Moreover, care must be 
taken that the crutch and the pendulum move exactly in 
the same plane, or the impulse of the pallets will tend to 
make the pendulum revolve in a sort of elliptical cone, in- 
stead of vibrating in a pkne. 

135. If the pendulum cannot be fixed to the wall, or if 
the clock frame itself is very strong and firmly fixed, it may 
be hung from a flat iron bar going over the top of the 
frame, and having a nick in the end broad enough to admt 
the chops. This bar will have to be screwed into both the 
back and front of the frame, and before the screw holes are 
made, it should be carefoUy adjusted, so that the pendulum 
will swing exactly in a plane at right angles to the pallet- 
arbor, and having the edge of the spring exactly opposite 


to the end of the pallet-arbor; which may be effected by 
taking out the arbor^ and looking through the two holes in 
the frame at the springs and seeing that it is in the middle 
and appears merely as a line. 


136. It is of nearly equal importance that the frame 
which carries the wheels should be strong and steady. The 
frame of astronomical and house clocks consists merely of 
two plates of brass joined together by pillars at the comers. 
The frames of turret clocks are made of iron bars so placed 
as to carry the arbors of the wheels. The old fashioned 
way is to make the frame consist of a front and a back set 
of bars joined by pillars like a house dock^ and requiring 
to be taken to pieces to get the wheels out when they 
want cleaning. This is still done in second-rate clocks^ 
though it may very easily be avoided. The frame ought to 
be so made that when once put together and screwed down 
to its standi the substantial parts of it should never be 
taken to pieces again ; and this is either done by making the 
bars which carry the wheels separate pieces^ screwed on 
to the more solid parts of the frame, or, what is better, as 
regards the smaller wheels, by making the brass bushes 
which contain the pivot holes to take out, and then the 
wheels can be removed separately, at any rate sufficiently 
to dean them. These loose bushes should have. their three 
screws set so as to form an isosceles, not an equilateral tri- 
angle ; for if they are at equal distances the bushes get put in 
in different positions, and as the pivot hole is not likely to be 
be exactly in the middle of all the screw holes, it gets dis- 


torted from its true place. If there are only two screurs 
there should also be one steady pin^ for the same pxirpose* 
The whole of the frame^ or of each side of it^ should be ca^ 
in one piece^ and the pillars should have broad shoulders and 
strong threads and nuts to unite the two sides together, and 
the frame should be from ^ to |th inch thick according to the 
size of the clock. Then the two barrels may be set in sepa- 
rate bars bolted to each side of the frame. All the rest of 
the wheels may be inserted by means of moveable bushes 
and without any extra bars to the frame. In this way the 
frame is strongs, as well as cheaper to make^ than if it is 
made in many pieces which have to be fitted to each other. 
But there are several varieties of clock frames which it may 
be as well to describe separately. 

187. First suppose there are to be four wheels in the 
train, as in a house dock, and all the arbors of the same 
length as the barrel, this being the most ordinary kind of 
construction. Then a frame such as this will be perhaps 
the best said most compact form. I have inserted the end 
of the bar which carries the pendulum, in order to show 
how it is to be placed if the pendulum is not fixed to the 
wall. I have not indeed seen a frame of this form> and 
possibly the arrangement n^y be found on trial to require 
some little modification; but a dockmaker to whom I sug* 
gested it said he should use it; and when it is necessary 
to resort to a four-wheded train, this kind of frame wiU 
carry the wheels with a less amount of iron work than usuaL 
F P P P are the four pillars, and a fifth might he inserted 
in a large clock in the cross bar which carries the scape- 
whed and the hammer lever, and which has the lo<ddng 


plate set upon a atud outside. The aumbera of the teeth 
which I have inserted aie of course arbitraiy, and may be 
altered in the proper proportions. 

I have put down the great wheel of the going part as 
turning twice more in the week thui the striking great 
wheel, because the striking part abnost always requires a 
heavier weight, and therefore a thicker rope than the going 


part. The requisite thickness of rope can be altered by 
altering the diameter of the barrel or the fall of the weight ; 
or if the fall is limited by the situation the weight can be 
increased and an additional pulley added. It is, however, 
to be remembered that every extra pulley adds very con- 
siderably to the friction and requires a still further addition 
to the weight to overcome that friction. I was lately able 
to take 401bs. off a striking weight by merely removing a 
guiding pulley which had been needlessly introdu^d. It 
is a matter of experiment what thickness of rope § given . 
weight hanging by a single moveable pulley requires : a 
good rope half an inch thick appears to be strong enough 
for any weight that is ever put to such a clock j and con- 
sequently the barrels need never be more than a foot long 
in a clock of this sort, as it is not supposed to be one of 
very large size. 

138. If the striking wheel has eight pins, it must have 
eight times as many teeth as there are leg^ves in the pinion 
of the wheel above it, in order that that wheel (which I 
called before the second wheel) may turn once for every 
blow that is struck (see § 93) ; and the second wheel must 
also drive the fly some exact number of times in each 
revolution of its own, since in turret clocks the fly arbor is 
made to carry the pin which is stopped by the detent D 
when they have done striking, as before described, the fly 
itself being set outside the frame and having arms a foot 
and a half long or more to carry the fans, and therefore 
moving much more slowly than in house clocks, where it is 
within the frame. These arms are not rigidly fixed to the 
arbor but to a socket which turns upon the arbor, and .has 


a ratchett and dick so placed that when the arbor is driven 
by the clock it carries the tLj with it^ but when the dock 
is stopped at the end of the striking the fly can go on 
by its own momentum^ as it conld not be stopped sud- 
denly without the risk of breaking something. It is this 
motion of the fly over the ratchett-teeth which makes the 
noise like winding up that may be heard in the tower after 
a church clock has done striking. 

In the above drawing the locking-plate or count-whed 
B L is outside the frame^ and is driven by a small pinion 
of 8 set on to the projecting arbor of the striking-whed. 
The lifting piece A G D is set on an arbor which is within the 
frame but projects through it and has the arm G B set on 
to it outside^ which works into the locking plate. At F 
there is a roller which rolls upon the circular plate with a 
piece cut out of it fixed to the second wheel. The snail at 
A on the centre whed raises that end of the lever, and 
therefore depresses the other end, so that the pin in the 
fly-stop slips past D and rests against E when the clock 
'gives warning.' When the snail lets the end A of the 
lever drop, the fly is set at liberty; and as the second 
whed begins to move, the circular plate and roller depress 
the lever so far as to let the fly-pin clear both the stops D 
and £, and then the locking plate by means of the arm 
B not only keeps the lever there but moves it a little 
fiEurther in the same direction, merdy for the purpose of 
taking the friction oS the second wheel which moves 
quickly, and transfering it to the locking-plate which moves 
very slowly. In fact the roller F is not strictly necessary, 
for the locking plate alone might be employed to raise the 

I 8 


lever^ as described in § 93 ; but it is safer to have it raised 
at first by a wheel that moves so quickly that the lifting- 
piece is sure to be out of the way before the fly has made 
one revolution. Of course the end A of the lever must be 
made a little heavier than the otiier end. 


139. I have supposed each train to have four wheels. 
But it is evidait that it will move with less friction if the 
teeth can be so proportioned that there may be only three 
wheels. Now if we assume that the centre wheel shall 
tarn once in the hour, we can hardly make the scape wheel, 
which is now to be the next to the centre wheel, turn in 
less than four minutes, which however can be done where 
there is a two seconds pendulum with a scape wheel 
having sixty pins instead of teeth in it; but it requires a 
large number of teeth in the centre wheel. In this case 
the great wheel may remain the same as before. In the 
striking train the second wheel is dispensed with, but a 
longer fly is required; and if the number of pins in the 
striking wheel remains the same the great wheel will also 
remain unaltered. The locking-lever however will have to 
be altogether worked by the locking-plate, which requires 
very accurate adjustment of the nicks to make the lever 
rise and fall exactly at the right time, and the locking-plate 
must be large in proportion to the length of the arm C B 
in order to raise it quickly enough. Altogether this is 
rather a critical movement, though I have seen it answer 
very well in well made clocks, such as Mr. YuUiam/s, who 
geueraUy uaea it, as wdl aa the correspondkg going train. 


140. But there is another way of making three-wheeled 
trains^ which is better oh several accounts. First of the 
going train : it is not at all necessary that the centre wheel 
should go round in an hour; for there may be a subsidiary 
wheel (i. e. a side wheel not forming part of the train)^ which 
turns in the hour^ and drives the dial-work ; such a wheel 
is in fact used in the three-wheeled train I have just now 
described^ to reverse the motion; and an extra wheel of this 
soit does not cause nearly so much friction as the introduc- 
tion of another wheel and pinion into the train. Suppose 
we make the great wheel to turn in three hours instead of 
six or eighty as in the four-wheeled train; then^ if it has 
120 teeth^ and the next pinion 12 pins or leaves^ the 
second wheel will turn in 18 minutes, and if that has 
108 teeth, and the scape-wheel pinion 12 as before, the 
same pendulum and a scape-wheel with 30 pins only, or 
an eight feet pendulum with 40, will do. This is the train 
in Mr. Dent's large turret clocks, with occasional variations 
of the numbers for particular purposes. 

The subsidiary wheel, to be worked by the great wheel, 
must be one-third of its size, so as to turn in the hour : the 
manner of fixing titis wheel I shall treat of under 'adjust- 
ing work.^ But if the barrel turns in three hours instead of 
eight, it will have to make 64 turns in the eight days 
instead of 24, and so will require a much greater length, 
and if all the arbors are made of the same length as this long 
barrel, it will make the whole machinery very heavy. In 
order to get rid of this objection, what is called a^ double 
frame is generally resorted to ; that is, the barrels and their 
great wheels have their back arbors in the same frame 


as the rest of the wheels, bat the front or winding ap 
arbors in a eepatate frame standing Home disttmce in front 
of the &ame that carries the other wheels. Bj this anange- 
ment the arbors of the other wheeb can be made lighter 
even than in a clock with a short barrel. 

There is another advantage in this aa regards the striking 
part, viz., that when the great wheel turns so often as once 
in three or even four hoots, the striking pins can be put on 
the great wheel itself, as it is large enough to carry the 26 
pins required for one-third of the twelve hours. The exact 
number is indeed immaterial, except that the mote pins it 
has the fewer coils of rope it will have to carry for the 
eight days. Now it is evident that the striking will be 
done with less waste of power when the power has not 
to be transmitted through the pinion of a second wheel, 
in which it loses considerably by friction ; I shall have to 



state afterwai'ds the surprising extent to which this loss of 
power takes place. The wheel next to the great wheel can 
then be made to raise the locking lever in the same way 
as the second wheel does in the four-wheeled train. The 
more turns the rope takes over the barrel the easier also it 
is to wind up^ and the difference in moderate sized clocks 
may be such as to save the necessity of having what is 
called a jack to wind up with, which is generally required 
for the striking part of clocks with short barrels, except 
with small bells. 

141. This apparatus may be either fixed or moveable. 
If it is fixed, as it ought to be when required at all, the end 
of the barrel opposite to the great wheel is made into a 
toothed wheel K (in the drawing of the four wheeled train), 
And by the side of it is put a smaller wheel or pinion J 
which works into it; and instead of the winding square 
being made on the arbor of the barrel, it is made on the 
arbor of this pinion, which of course gives greater power 
in winding the clock up, by requiring more turns of the 
key or winch to do it. The pinion is sometimes made to 
push backward and out of gear with the wheel when the 
clock is not being wound up. The loose jack consists of 
a frame contaming a wheel and pinion with the arbor of 
the pinion squared to form the winding square : the arbor 
of the wheel is a socket with a square hole in it which fits 
on to the barrel arbor ; so that when the jack is applied 
and made to rest on some part of the clock frame, it be- 
comes (except that it requires a thicker arbor as will be 
explained presently) just the same as if the wheel had been 
set on the barrel itself, and the pinion in the clock-frame. 



The going part never requires a jack wheel to wind it 
up except in veiy large clocks : but I am not sure that it 
would not be better to apply such a wheel (I mean of 
course a fixed one) more frequently, even where the weight 
is not too great for a man to wind up without it ; because 
the effect of having to wind up on the barrel arbor is that 
it must be thick enough to carry a winding square strong 
enough not to twist, and therefore that pivot requires to be 
a good deal larger, and therefore has more friction, than 
would be necessary if it had merely to carry the weight. 
The two cast iron wheels that require to be added for 
winding by a jack wheel cost very little, and the plan also 
saves something, in the construction of the barrel arbor 
and great wheel. By a little contrivance one winding 
pinion might be made to apply to all the three barrel 
wheels : indeed there is no reason why the winder itself 
should not have that pinion, instead of a pipe, set on an 
arbor, fitting into proper holes to wind up all the barrels by. 

142. Loi% barrels however cannot always be properly 
used; because if the rope does not hang down, but is 
carried over a fixed pulley above or at the side of the dock, 
and there is not a considerable distance from the barrel 
to the fixed pulley, the rope wiU not run straight enough 
off the barrel towards the end, but obliquely as shown in 
the last drawing, and consequently, instead of travelling 
along the barrel as the clock is wound up, it will turn back 
again and overlap itself; and even if it does wind up without 
overlapping, there wiU be so much rubbing between the 
successive coils of the rope as will soon wear it out, and 
will moreover render a heavier weight necessary. In all 


cases where a fixed pulley is required^ it ought to be as far 
as it can be placed from the clocks in order to diminish the 
obliquity of the rope as much as possible ; I should say 
tljat the distance ought not to be less than eight times the 
length of the barrel 

Sometimes in order to avoid long barrels the rope is 
made to go twice over them; but this is a most abominable 
practice; for the weight has of course more power when 
the rope is coiled the second time over the barrel^ as that is 
the same thing as making the radius of the barrel larger 
by the thickness of the rope, and so the force upon the 
dock is not uniform; and secondly, it wears out the rope 
more than anything else that can be done to it. When 
the position of the clock is such that the fixed pulleys 
cannot be as far off as about eight times the length of the 
barrels, the clockmaker should not be allowed to use long 

143. There are however two ways by which the advantages 
of the long barrels may be obtained without their disadvan- 
tages. The first is by using wire ropes instead of hemp j* 
but great care must be taken that they are good, as there 
are some very bad ones made ; and they should be tried for 
some time beforehand with a heavier- weight than they will 
have to bear, and above all by winding them up frequently, 
or they will cause constant trouble smd expense after the 
dock is put up. In this way very elegant clocks can be 

* In the Windsor Castle clock Mr. Vulliamy used catgut ropes 
(vire ropes were not then invented) ; but they are enormously ex- 
pensive : he was told that 17,000 theep contributed their entrails to 
compose these catgut lines. 


made^ ¥dth great strength in a small compass^ and with 
only three wheels in each train^ and striking from the great 
wheel. A picture of one is given in the frontispiece. 

144. The second method is to put about twenty striking 
pins on one side of the great wheel and another set half 
way between them on the other side^ and make them raise 
two hammers alternately: forty^ or even thirty^ are too 
many for one side of any but a very large wheel made in a 
particular way^ which will be described presently. Mr. 
Dent, until he lately at my suggestion adopted this other 
method, constantly used the two hammer plan, and it 
answers very well ; and such clocks strike with much less 
waste of power, even when made in cast iron, than the most 
highly finished clocks with polished steel pins set on the 
second wheel. A wheel of this sort with forty pins will 
turn very nearly twice m the twelve hours, or thirfcy-one 
times in the eight days, for which a barrel of moderate 
length will be sufiB.cient, especially as a smaller weight and 
therefore a thinner rope will do the work than in a train 
striking from the second wheel. 

145. Or again, cheap clocks with short barrels might 
be made with about twenty pins on the great wheel and one 
hammer, to wind up twice a-week : a condition to which a 
good many clocks are practically reduced which pretend to 
be eight-day clocks, from being so placed that they will not 
go fuU seven days. And those who can only afford to 
spend a small sum upon a church clock had much better 
have a good one made in this way than a bad one to go 
eight days. 

In every clock the going part ought to go half a day 


longer than the striking part, in order that, if it is for- 
gotten, the clock may proclaim by its silence that it wants 
winding up before any harm is done. 

Lastly, the reader may be reminded that in calculating 
the teeth and pinions required for driving the hour wheel 
by the great wheel separately from the train, the number of 
the teeth of the great wheel is immaterial ; for in that case 
the hour wheel will evidently turn in the same time as if it 
were put in the place of the great wheel. Suppose the 
scape wheel turns in m minutes and its pinion has j9, leaves, 
and let ti be the number of teeth of the wheel that drives 
it, whose pinion has /?, leaves, and let ^ be the number of 

teeth of the hour wheel : then :~^ must = 22 . 

Suppose j?i = 8 and jOa=10 (which will be enough if 
they are lantern pinions, though not otherwise) and m=2 ; 
then if ^s=24, ^, must =100, and the second wheel will 
turn in 25 minutes ; and if t^, the number of teeth of the 
great wheel,=l44, it will turn in six hours, or will require 
only thirty-two coils of rope for eight days. In this case 
the hour wheel had better be made as a large lantern 
pinion. And if the great striking wheel has twenty pins 
or cams on each side raising two hammers alternately, as 
above mentioned, an eight-day clock may thus be made 
with only three wheels in each train, with moderate num- 
bers for the teeth, and without requiring the scape wheel 
to turn so slowly as in four minutes, and without resorting 
either to long barrels or wire ropes. 

146. When the great wheel is the striking wheel, the 
second wheel must turn exactly one-half or one-third round 
for every stroke, and the circular plate before described in 


the four wheeled train must have two or three pieces in- 
stead of one cut out accordingly, to let the roller on the 
locking lever fall at the proper time. Moreover this wheel 
will turn too quickly for it to work the locking-plate by a 
wheel and pinion, as it would require at least 234 teeth on 
the locking-plate wheel. If the clock is wound up by 
a jack wheel as before described, the barrel will ride upon 
the arbor of the great wheel, and that arbor may carry a 
pinion or small wheel to drive the locking-plate, just as 
the striking wheel of the four wheeled train did. Sut 
when the weight is small enough to be wound up without 
a jack, the arbor must belong to the barrel and not to the 
wheel, and therefore cannot drive the locking-plate. Jh 
that case it must be done in one of the following ways. 

The most simple method is to set the pinion on the 
great wheel itself : and in every case the wjieel or pinioii 
that drives the locking-plate, if it is set on the striking 
wheel, must have as many teeth as there are striking pins, 
assumiag the locking-plate to have 78, or in the same 
proportion. 2. Although the second wheel turns too 
quickly to drive the locking-plate by a toothed wheel and 
pinion, yet it may do so by means of a ratchett-whed on 
the locking-plate and two gathering pallets, (like the com- 
mon repeating movement, § 90) set on the arbor of the 
second wheel: the locking-plate must also have a click set 
over the ratchett to keep it in. its place when the pallets 
are not acting. A third way is to make the barrel arbor 
hoUow nearly up to the end which forms the winding 
square, and make the arbor of the great wheel go into it; 
so that the barrel arbor comes through the front frame. 


but the wheel arbor goes the back frame and carries the 
pinion that drives the locking-jdate. This^ when it is well 
done^ is a remarkably neat arrangement^ but it requires very 
good machinery to bore the hole truly all down the barrel 
arbor, and if it is not done truly it had better not be done 
at all, as it is sure to stick fast. These short trained clocks 
may of course be put in a frame of the same kind as the 
four wheeled one of which I gave a drawing. But they 
afford facilities for a more convenient arrangement, by 
setting the going and striking parts, or at least the upper 
wheels of them, in separate small frames, which screw on 
to a large and strong horizontal frame cast all in one piece, 
which may carry the great wheel of the striking part, and 
indeed of the going part also. 

The clock in the frontispiece is an example of this kind 
of frame. The great striking wheel is carried by a large 
cast iron cock on each side, bolted to the main horizontal 
frame, the surfaces in contact being planed so as to fit 
accurately. The second wheel and the fly arbor are carried 
by a small triangular frame at each end, which screws on 
to the main frame. The locking-plate turns on a pin or 
stud set on the frame, and as there is a jack wheel to 
wind up by, the great wheel arbor comes through the front 
cock and drives the locking-plate. The great wheel of the 
going part being lighter and smaller than the striking 
wheel is carried in the same triangukr frame as the rest of 
the going wheels ; but tiie scape- wheel and the bushes of 
the second wheel are made to take out without removing 
the great wheel or taking the frame to pieces. It is not of 
much consequence whether the second wheel can be got 


completdj out, or not^ for if the bushes are removed it 
can be cleaned within the frame as well as if it was taken 
out. The scape- wheel however^ and the pallet arbor, should 
be made to take out completely, without undoing the frame ; 
and for this purpose nothing is required but the bushes to 
be fixed with screws instead of rivetted in, as there is always 
room enough for the scape wheel to pass the second wheel. 
147. I proceed to point out some other things which 
require to be attended to in turret clocks, in addition to 
those rehkting to clocks in general, mentioned in the first 


The scape-wheel and its arbor, as before observed^ 
should be as light as it can be made of proper strength j 
remembering also that the fewer pins or teeth it has tho 
lighter it should be made, as it has to move farther at each 
beat. The pivot of the pallet arbor ought to come close 
up to the pendulum spring ; for if it does not, the force is 
communicated to the pendulum with a tendency to twist 
the crutch : no such effect will of course be visible, but the 
tendency to it produces an unnecessary friction on the 
pivot, and requires greater thickness and therefore weight 
in the crutch to resist it. In the Meanwood clock the 
crutch is entirely outside the frame, straight, thin, and 
broad, and therefore as light as possible, the pivot being 
set in a cock like that of a common house clock. Some- 
times however the frame is so made that the pallets and 
crutch could not be taken out if they were set in this way;. 


and in that case the pivot should be set in a bnsh projecting 
as far as the pendulum springs the crutch having a bend^ 
and springiQg from within the frame in the usual way. It 
is to be remembered that the drutch is itself a pendulum 
of which aU the weight is carried by the pivot holes, and a 
heavy crutch and arbor produces considerable friction, 
which, like all other friction, is variable, and moreover con- 
sumes force, diminishing the arc of the pendulum like the 
dead friction on the pallets, of which the effect has been 
ahready explained. 

Eor the same reason the pivots ought to be as small as 
they can safely be made, and also with as little shake as 
will let them move freely; and the crutch should be long 
that little may be lost by the necessary shake of the 
pendulum in the fork. Whether the crutch ends in a 
fork embracing the pendulum rod or a pin going into a 
hole in it, it should be made with some screw adjustment 
to put the clock in beat. And in trying whether a clock is 
in beat the pendulum should be allowed to move only just 
far enough to let it escape, and then the smallest deviation 
will be easily observed. It may be a convenient rule for 
unprofessional clockmakers to remember, that in all cases 
if the right hand beat {i.e. the beat heard when the pen- 
dulum is at the right) comes too soon after the left hand 
one, the fork wants moving to the right with reference to 
the crutch, and vice versd. 

148. A very neat contrivance has lately been introduced 
by Mr. Dent into his turret clocks. It occasibnally happens 
that if the pendulum is moved while the scape- wheel is 
standing still without any force upon it, the end of one of 


the pallets just catches the point of atooth so that it cannot 
slide up the pallet, and the momentum of a heavy pen- 
dulum is then certain to break the tooth. In order to 
prevent this the two sides or prongs of the fork are set 
upon springs, which tend to bring them together and ke^ 
them close to the pendulum rod, with a stop of the proper 
thickness between them. The springs are strong enough to 
give the impulse to the pendulum, but if one of the pallets 
should be stopped, the spring on the opposite side of the fork 
gives way and lets the pendulum go on. The reader may 
fancy that a single spring in the crutch would do as well; 
but it would not; for a single spring would not have suffi. 
dent resistance boti wayB to communicate the impulse to the 
pendulum without bending; therefore there must be two 
springs, both in a state of tension in opposite directions. 
The little frame that carries the springs screws on to the 
crutch in the same way as a common adjustable fork. 


149. In turret clocks the dials and dial work aie ge- 
nerally at some distance firom the clock. The dial work 
consists of an arbor which carries the long hand with a 
wheel upon it, and upon this arbor is the short hand socket 
with its wheel, and the two work together in the way 
before described in house clocks. These wheels are all set 
in a small separate frame fixed behind the dial, and all the 
wheds and arbors should be made of brass, as it is gene- 
rally not very easy to get at the dial work to oil it ; and 
so if made of steel the pinions are liable to rust. The 


hour arbor projects througli the front of this frame^ with a 
half universal joint, and if it is on the same level as the 
clock, it is merely connected with the hour wheel of the 
clock by a straight rod. If the dial work is above the 
clock it must be connected by bevelled wheels, one pair at 
the clock and the other pair at the dial work, or, if there 
are more dials than one, at some convenient point where 
the rods converge from the dial work of each dial, acoasding 
to the position of the clock. These vertical and horizon- 
tal leading-off rods are sometimes made of tubing, either 
iron or brass, as it is stronger than a solid rod of the same 
weight and strong enough to resist bending by its own 
weight as well as twisting. In the same way I may 
mention that small iron gas-pipes with solid pivots fixed 
into their ends are sometimes used when cranks with long 
arbors are required for connecting the hammer with the 
wire coming up from the dock. But where the vertical 
rod is required to be long and consequently heavy, it is 
generally better to make a wheel at the top of it rest on 
friction rollers, and then the weight of the rod has no ten- 
dency to bend it, and it only requires to be strong enough 
not to twist, for which a half inch iron rod is sufficient in 
almost any case. These friction rollers may be made in 
various ways ; on the whole the best way appears to be the 
following, which will be readily understood from the 
drawing over the page. 

The rod has a circular plate fixed to its top, and there 
is a similar plate fixed to the beam through which the rod 
passes. These plates are not quite flat, but they each have 
a broad circular groove in them near their edge and the 


grooves exactly face each other. 
Now if tliree or four equal balls 
or cheese-shaped rollers were laid 
on the groove in the tower plate, 
and the upper platelet down upon 
them, it would turn with no fric- 
tion except that of the roUinff of 
the balls, which would alwajri 
keep their original distance. The revolving roof of the 
building which contains the great telescope at Cambridge 
tarns upon cannon balls in this way ; and the balls run- 
ning in the grooves cause it always to turn on the same 
imaginary axis without the necessity of a real one. How- 
ever on this small scale, in order to prevent the balls from 
shpping uid so getting all together, they are set upon a 
three legged axis with a hole in the middle which surrounds 
Uie rod ; but it has nothing to do as an axis : that is, it 
curies no weight and has no contact with the rod, and 
has merely to prevent any of the balls from not rolling on 
in case the upper plate should for a moment be lifted off 
that baU. Consequently the friction (so far as regards the 
weight of the rod work) is true rolling friction, which is 
much less than the friction of friction v>heelt of any mode- 
rat« size carrying the weight upon their axes. I under- 
stand that rollers of this kind have been applied to huge 
weather-cocks with long rods to work a dial within a 
building, and tried with heavy additional weights without 
producing any material difference in their freedom of action. 
It should be observed that the cross section of the rollers 
must be of higher corvatnre than that of the grooves, so 


that they may only have contact at one point, and not nib 
as well as roll : complete surface contact without rubbing 
could only be obtained by conical rollers between conical 
plates, the points of all the cones converging in the centre of 
the rod ; but such rollers would have a tendency to slip 
outwards, and would require greater accuracy in making 
than these ' cheese shaped' rollers. 

The long vertical rod is connected with the arbor of the 
bevelled wheel in the clock by means of a half universal 
joint; and in fixing it in cold weather it should be remem- 
bered that an iron rod thirty feet long will become ^th of 
an inch longer with 40^ of additional heat, and brass tubes 
about half as much more ; and play enough must be left 
in the joint accordingly. 

Where the clock does not stand in the middle of the 
chamber, under the point of convergence of the dial rod?, 
but near the wall, an additional pair of bevelled wheels is 
required. This might be avoided by making bevelle4 
wheels of a different angle from 45^, so as to lead directly 
to the two side dials from the place where the vertical rod 
comes up ; but nothing material either in expense or fric- 
tion would be saved thereby. The bevelled wheels, espe- 
cially those which have to work several others, ought to be 
large and strong; and when, as sometimes happens, their 
diameter is limited by want of room in the clock, it ought 
to be made up in thickness, which however gives more 
friction than krger wheels with more teeth. 

The adjusting work, for setting the hands when the 



clock wants altering, is sometisEies very defectiye. I have 
seen a clock in a cathedral put up only a few years ago, 
as I was told, by a London firm of large business, in which 
the only means of putting the clock forward is by taking 
oat the bush of one of the wheels (which is famished with 
hand-screws for the purpose) and shifting the place of the 
pinion upon the teeth of the next wheel. I should think 
the man who has the charge of such a clock takes very 
good care always to keep it on the gaining side, as he can 
stop it for a few minutes every now and then with in- 
finitely less trouble than he can put it forward in this 
barbarous fashion; The adjustment is generally made 
merely by a friction spring as in a house-clock (73). And 
where there is only one small dial this method is safe 
enough ; but otherwise the spring must either be so strong 
that the hands cannot be altered without applying great 
force to the wheeb, or else they will be liable to slip for 
want of sufficient friction or pressure of the spring. In 
large clocks therefore a better way of making the adjust- 
ment is by what may be called a square ratchett, i.e, one 
which wants the click lifting by hand to enable it to pass 
either way, each division corresponding exactly to one or 
two minutes ; or by clamping screws. 

When there is no hour wheel in the train this ratchett 
is sometimes set on the great wheel arbor, turning in three 
hours suppose, and with a large bevelled wheel attached to 
it, which drives the first leading off bevelled wheel, of one 
third its size and number of teeth. This has one advantage, 
yiz : that the friction of a subsidiary hour-wheel is 
avoided ; which however is much less than if it were an 


additional wheel and pinion in the train : the large bevelled 
wheel also carries three pins instead of one, to let off the 
striking part* But it has several disadvantages : first, the 
small bevelled wheel is necessarily of small dimensions; 
since even the large one, which is three times the size, 
must be made smaller that the great wheel itself, or it 
could not be got into the frame : secondly, the discharge 
ing pins cannot be relied on to let off the striking part 
correctly within several seconds in this way : thirdly, when 
the leading off is wanted in any but a vertical direc* 
tion, the arrangement becomes clumsy and inconvenient ; 
fourthly, the ratchett-wheel, bevelled wheel, and pins add 
nearly two inches to the length of all the arbors; and 
fifthly, in some positions of the wheel the adjusting click 
cannot be got hold of without great difficulty. I think 
any of the following methods are better, and some of them 
may be easily adopted in any clock, according to circum* 

When the hour-wheel is not in ihe train it is to be a 
thick wheel of the proper size to be driven by the great 
wheel according to its velocity. It is not fixed to its arbor, 
but rides upon it close to a plain or blank wheel, which is 
fixed to the arbor, as are the discharging snail and the first 
leading off wheel, if any are required. If the dial is a small 
one the hour-wheel and the blank-wheel may be connected 
by a friction spring as usual ; otherwise the connection may 
be made, either by one or two clamping screws, or by a pin 
which goes through a hole in the hour-wheel and into any one 
of sixty or thirty holes in the blank wheel, being kept in its 
place by a spring, and having a projecting handle by which 



it can be pulled out and the relative position of the two 
wheels shifted through a space corresponding to one or two 
minutes. Another waj^ perhaps still more simple, of 
making the adjustment by a definite quantity, would be to 
make the hour-wheel^ instead of riding or turning on its 
arbor^ to slide upon it by means of a square^ or what is 
called a hey; when you want to alter the clock you would 
only have to slide the hour-wheel out of gear with the 
great wheel and slide it in again with different teeth in 
contact^ and it might be kept in its place either by a pin 
or a spiral spring of wire round the arbor. Where the 
clock has a train remontoire, which may itself require 
adjusting^ as described in § 177, clamping screws must be 
used ; because altering the remontoire- by any less quantity 
than the twenty or thirty seconds at which it lets off would 
make the hands point wrongs if they could not be altered 
by as small a quantity as the scape-wheel. Where either 
of the other methods is adopted^ so that the hands cannot 
be altered by less than a minute^ the smaller adjustments 
must be made by stopping the scape- wheel for the requisite 
number of seconds ; and on this point I shall have to make 
some farther remarks in § 187. 


151. The size and strength required for the going part 
of a clock depends entirely upon the number^ size, and 
situation of the dials ; though there seems to be a notion 
among clockmakers that the going part and the striking 
part ought to correspond in size^ and I have seen a dock 
without any dials^ in which the going train was heavy 

DIALS. 197 

enough to work four dials 10 feet wide. They ought to 
be treated quite distinctly; though of course it will fre- 
quently happen that where there are large dials there will 
be a large bell also. The more turns in the week the great 
wheel makes^ the less strength and size it obviously re- 
quires ; and in all cases broadness of the teeth^ i. e. thickness 
of the wheels^ should be looked to rather than depth of the 
rim, as broad teeth cut the pinions less than narrow ones. 

The size of the intended clock dials is a matter which 
church architects frequently pay no attention to until it is 
too late, or do not understand. Mr. Yulliamy states in his 
pamphlet the sizes of several well known pubUc dials in 
London : 

St. MartinVin-the-Fields - - - 8 ft. 

St. James's, Piccadilly ... 10 ft. 

Islington church - - - - 9 ffc. 

The clock on the Queen's Stables - 6 ffc. 10 in. 

St. Paul's 17 ft. 

Horse Guards - - - - • 7 ft. 5 in. 
To which I add from other information : 


St. Luke's, Chelsea* - 6 ffc. 10 in. 72 ffc. 

Bow Church - - - 9 ft. 70 ffc. 

Marylebone Church - 7 ft. about 60 ffc. 

Koyal Exchange - - 9 ffc. 90 ffc. 

* As a proof of the reliance to be placed on architects in these 
matters, 1 have seen a letter from an architect of some reputation, 
stating as a justification for the dials which he had prepared, and 
which the clockmaker objecte4 to, that the first two dials in this list, 
and that of the Horse Guards, were only 5^ feet in diameter : which 
his employers had of course believed. 


And the great clock at Westminster is intended to have dials 
23 feet wide, at the height of about 220 feet. 

The result evidently is, that unless all the above 
dials are too large, which anybody may see they are not, 
dials ought to be so placed that they can be about a 
foot in ^idth for every ten feet from the ground; though 
tliis is not sufficient for heights under 50 feet. Any one 
who looks at dials of less than this proportion, such as 
St. Pancras, 6 ft. 6 in. at a height of about 100 ft., or the 
dial on the church in Chester Square, which is a little more 
than 4 feet, and at a very moderate height, will see a fur- 
ther proof of the necessity, not only of making the dials 
large enough according to the above rule, but of placing 
them where they can be made large enough for their own 
height, without being too large for the surrounding parts 
of the building. The consequence of not attending to 
this is that the tower is defaced, and its ^ details^ over- 
whelmed, by what appears at a Httle distance only a great 
black spot, too large for the building and too small for the 
clock. It is hopeless to make a clock face an archi- 
tectural ornament : at least every attempt that I have seen 
of that sort in a gothic building ha« Jn a most wretched 
failure, both in architectural beauty and horological dis- 
tinctness. The best thing that can be done is to put the 
dials on some plain flat surface large enough for the pur- 
pose; and if such a place is not provided at first, it is ten 
to one that a clock-face will some day be substituted for 
the tracery of a window (as three illuminated dials have 
been in the heads of the windows of the fine tower of St. 
Mar/s at Beverley), or buUt round the bottom of the 

DIALS. 199 

spire^ projecting ' like the eyes of a great cod-fisb/ as I 
heard remarked by a spectator of three dials stuck round 
the spire of a small country church by the architect just 
now alluded to. 

152. The material of the dial may be stone^ slate^ 
copper^ or iron. If the dial is made of stone it should 
have the part within the figures, in which the short hand 
traverses, cut out or countersunk to the depth of an inch 
and a half, in order that the long hand may lie closer to 
the figures, for the sake of avoiding pa/rallax as much as 
possible. The efiect of parallax is that when the hand is 
in any position except nearly vertical, the line of vision 
from the eye of the spectator to the hand of the clock does 
not fall on the place to which the hand is really pointing, 
but somewhere above it, depending on the distance of the 
hand from the face and of the height of the clock above 
the spectator. The same may be done with slate dials; 
but they are more expensive to cut than stone. Stone 
dials should be painted all over, black for gilt hands, or 
white for black hands, like one of the Horse Guards dials. 
Slate may also be painted, though it will keep a tolerably 
dark colour without painting, especially if it be occasionally 
oiled. Por large dials the slate will have to be in two 
pieces. The figures and minute marks are better cut than 
merely painted and gilt on the flat surface, as their place 
is then fixed once for all : otherwise they are not unlikely 
to be incorrectly divided in subsequent painting* In 
copper dials this cannot be done, nor the countersinking of 
the middle of the dial, without great additional expense ; 
and the dial is also obliged to be made convex in order to 


preserve its shape: the convexitj should however be as 
little as possible^ as it either increases the parallax or re- 
quires the long hand to be bent to come nearer the minute 
marks^ which has a bad appearance : an inch in 5 feet seems 
to be enough convexity. 

I am not aware that there is any objection to cast-iron 
dials, provided they are kept weU painted; and they are 
cheaper than copper. They have also the advantage of re- 
quiring no convexity, and the centre can be cast counter- 
sunk. I have seen the figures and minute marks made pro- 
jecting, which however does not look weU, and is objection- 
able, as the hand may be blown against them. They are better 
countersunk a little, for the reason I gave just now; and I 
may observe that the countersinking should not be square, 
so as to leave comers in which the wet will lodge, but a 
curved hollow like the fluting of a pillar. Bound dials 
always look better than square ones with the spandrils 
filled up with some attempt at decoration : a dial is from 
its nature necessarily a round thing, and it has no busi- 
ness to pretend to be a square one; even arches set in 
square heads only belong to the worst style of gothic archi- 
tecture, and round windows belong to the best style. But 
dials should not be set deep in the wall, like windows, or 
the rain will not wash them. 


153. Illuminated dials are made of an iron &ame work 
or skeleton, in the form of a ring, consisting of the figures 
and minutes and the three rims which bound them. The 

DIALS. 201 

middle of the dial must be one piece of ground plate glas9, 
for any division in it would cast a shadow and be mistaken 
for the hands. Glass is also put behiud or between the 
figures. Several gas-burners are put behind the dial^ and 
the cocks of the burners are connected by levers with the 
dial work^ so that the clock itself turns the gas nearly off 
(but not so far as to put it out) when the day dawns^ and 
turns it full on when it becomes dark. This is adapted to 
the different lengths of the day by pins which are screwed 
in from time to time by the person who has the care of the 
clock ; though of course by the addition of more compli- 
cated machinery the regulation of the 'gas-movement' fot 
the length of the day might be made self-acting. 

Another way of illuminatiug a dial^ and a much better 
one^ when the building admits of it^ is that which is used 
for the Horse Guards clock; on which a strong light is 
thrown from a lamp^ with a reflector^ placed on the pro- 
jecting roof in front of the clock tower. 


154. In nearly all public dials the figures are made too 
large; for the larger they are, the more they contract the 
really useful part of the dial; it will be seen that in most 
public dials the figures nearly touch each at their inner 
circumference; and consequently that part of the long 
hand which is over the figures cannot be distinguished at 
all at a moderate distance, and the dial might as well be 
only two-thirds of its actual size. Nobody wants to read 
the figures ; twelve large spots would do just as well,— and 



better; for then the long hand could be more dearly seen 
between them. The figures ought not in any case to 
be larger than a quarter of the radius of the dial; and 
they ought to be rather narrow^ or their different strokes 
close together^ so as to keep the VII and YIII and the 
other wide %ures visibly apart from each other. It may 
be as well to inform those who have not seen it tried^ 
that it has a bad effect to make the figures narrower at 
their inner circumference than at the outer^ as if the strokes 
were formed by radii of the dial ; and this is a further 
reason why the figures should be small. The five-minute 
marks should be a good deal larger than the other minate 
marks^ or it is not easy to distinguish which minute belongs 
to the figures that consist of several strokes. 

155. Both the hands should be broad; and the short 


hand only should have a heart or a broad part a little way 
from its pointy and the point of it ought to be entirely 
within the figures, and not haH covering them as it some- 
times does : the long hand should be straight, plain, and 
ending in a point, like a straight sword, just half way 
over the minute marks. These directions may seem need- 
lessly particular ; but the object of clock faces is to show 
the time distinctly as far as possible; and any one who 
will compare the few in which these things are attended to 
with the many in which they are not (indeed I hardly 
know any in which they are all attended to) will see that 
the attention is not thrown away; and as it costs no more 
to make the figures and hands in this way than in the com- 
mon way, there is no excuse for not doing it, unless some 
other way really better can be found. The material of the 

HANDS. 208 

hands is almost always copper^ gilt^ and stiffened at the 
back for some distance with a rib of brass. Ironwonldbe 
lighter for the same strength^ and when zinced or ^gal- 
vanized^ before gilding there appears to be no reason why 
it should not answer as well as copper : indeed it is pro- 
posed to make the hands for the great Westminster clock 
in this way. 

156. There is some difference of opinion whether the 
hands should be counterpoised externally or internally ; 
for one or the other they must be on account of their great 
weight. If they are counterpoised internally^ the force of 
the wind is not counterpoised at all, and it tends to bend 
the hand and drive it against the dial, or, if the hand is 
strong enough to resist that, to bend the arbor which 
carries it; whereas if it is counterpoised outside, the pres- 
sure of the wind against one arm of the lever is balanced 
by that against the other, and there is no lateral pressure 
on the arbor. And in like manner when the wind tends to 
htm the hand in one direction it will tend to turn the 
counterpoise in the other direction, and so there will be no 
strain on the teeth of the wheels, which in large dials, with 
the leading off or dial wheels not strong enough, has been 
known to break them. The objection to an external coun- 
terpoise for the long hand is that if it is gilt it will be 
mistaken for one of the hands, and if it is black like the 
£ace it will partially hide the hour-hand for about twx) 
minutes in every hour. My own opinion is, though it is 
contrary to the modem practice (which has perhaps been 
adopted by clockmakers to guard against the propensity of 
painters and gilders to gild the counterpoise), that the 


partial hiding of the short-hand for a minute or two is of 
far less consequence than the pressure upon the arbors and 
upon the train of the clocks which is caused by the want 
of an external counterpoise^ whenever there is a high wind ; 
and I should accordingly (especially where the dial is high 
and large) have a black external counterpoise to the long 
hand^ with its broad part (which it requires to make up 
for the want of length) falling just within the inner rim of 
the figures^ and therefore just beyond the heart of the 
short-hand. The black counterpoise of the short-hand will 
of course cover nothing but the black face. There ought 
however to be set on each of the dial rods inside^ a longish 
arm in the same way as an internal connterpoise, to enable 
a person fixing the dial work to know the position of the 
hands outside. It is not uncommon to see tliree or four 
dials on a tower all showing different time^ or to hear the 
clock strike when the minute-hand is not pointing to 
the 60th minute. When internal counterpoises are used, 
the arms which carry them should be as long as the space 
will allow, because the shorter the arm is which carries the 
weight the heavier the weight must be, and the greater the 
constant pressure on the arbors and sockets. 


157. As turret clocks take longer time to wind up than 
house clocks, they still more require some maintaining 
power to keep them going during that time. In the best 
clocks Harrison's going ratchett (7^) is used; but care 
must be taken that the spring has play enough to keep the 


tram going with proper force the whole time of winding. 
In clocks of moderate size the spring may very con- 
veniently be made merely of steel wire^ wrapped in coils 
with about }th of an inch interval between them^ round a 
lod bent into in a circular arc^ of which one end is fixed 
to a spoke of the great wheel and the other end runs 
through a socket screwed to a spoke of the ratchett wheel, 
so that the spiral spring is compressed between the two, 
being kept in its place by the circular rod run through it. 
In very large clocks however it is difficult to get force 
eEOUgh with springs of this kind, and therefore they must 
be of the shape drawn in § 72; and in all cases there 
should be two springs on opposite spokes, both for greater 
security and because you can get more play with two than 
with one of twice the strength. 

158. But there is a more common and much cheaper 
kind of maintaining power for turret clocks, which goes by 
the name of the boU<md ahuUer; in which a weight at the 
end of a lever is made to drive one of the wheels in the 
train for a few minutes. The drawing at p. 209, though it 
is not intended to represent the common construction, wiU 
serve to explain it. The lever A £ D carrying a weight 
at its end turns on an arbor A, and at B there is a bolt or 
click (shown as a fixed bolt in the drawing), which wiU 
allow the lever to be raised, but not to fall again without 
turning the great wheel with it. This bolt is more com- 
monly made to slide in a socket, like the spring bolt of a 
door; but this is a bad plan, as it is very liable to stick, 
and then the clock is left without any maintaimng power. 
On the long arm of the lever there is placed a cap or 


shutter (different from that shown in the drawing) which 
covers the winding square C when the lever is down or not 
in action ; and therefore the man has to raise the lever and 
shutter out of the way^ before he can put the winder on. 
The bolt is generally made to act on the centre wheel be- 
cause that requires a less weight ; but where the dial work is 
driven independently by the great wheels as in § 150^ this 
is objectionable^ because it causes the centre wheel pinion 
to act backwards : if it is inconvenient to apply it to the 
great wheel it would in this case be better to apply it to 
the hour-wheel which is driven by the great wheel. And 
in any case it is to be remembered that the maintaining 
power only acts until it has run the bolt out of gear^ and 
drops on to the stop G, or some convenient part of the 
clock &ame^ and it will of course run itself sooner out of 
a wheel that turns in an hour^ and stiU more out of one 
that turns in about twenty minutes, than it will out of the 
great wheel. It might indeed be made to act longer by 
setting the lever upon an arbor concentric with that of the 
wheel or nearly so, and making the dick throw itself out 
by coming against a pin in the frame when it had got low 
enough; but the more simple and safer way is to make it 
act on the great wheel as the dock weight does. 

But the great defect of all the common methods of 
constructing the bolt and shutter is that you have no se- 
curity for the lever being raised far enough to keep it in 
action during the whole time of windiQg up, especially if 
the man loiters over it, as he is very likely to do in winding 
a large dock with a heavy weight. There is also a smaller 
defect of just the opposite kind, viz : that the maintaining 


power generally remains in action for some time after the 
winding is done^ and so there is a double force on the 
clock. This does not happen with a spring going barrel; 
and though that may also theoretically run itself down if 
the man is too long winding, yet practically it can hardly 
happen, because if he stops to rest he will of course let 
the weight go oflf his own hands on to the clock, which 
win at once restore the tension of the spring, and he will 
begin winding again with the same maintaining power as 
at first. Indeed in a clock that takes several minutes to 
wind and has a spring going barrel, it is better to direct 
the man always to stop and leave hold of the winder in 
the middle of the winding. 

But, on the other hand, it is impossible to make the 
spring act with equal force during the whole time. In 
clocks of the common construction this is of very Httle 
consequence, for the spring is only required just to keep 
the scape-wheel going, as a heavy pendulum wiU go for 
many minutes without any sensible variation, even if it 
receives no impulse at all &om the scape-wheel. But in a 
dock with a remontoire in the train, which always requires 
a certain amount of force to Uft or wind it up, it is evident, 
that if the spring is to be strong enough to do it when it 
is nearly run down, it must act much more strongly than 
is necessary at first; and the larger the clock, or the longer 
it takes to wind up, the greater must be the excess of force 
to be left constantly on the clock train, in order to bring 
the spring to the requisite tension; which, though it will 
not reach the escapement, being intercepted by the remon- 
toire, is of course a defect, and helps to wear out the dock; 


and therefore the spring-going barrel is not so well adapted 
for a remontoire clock as some kind of maintaining power 
which acts by gravity, and so is pretty nearly constant. 
Mr. Airfs going-barrel, which will be described in § 160 
effects this object; but it so enormously increases the ex- 
pense and trouble of making the dock, and is so unfit tp 
leave in the hands of any but skilful persons afterwards, 
that it is impossible that it can ever be generally used. I 
have therefore attempted to contrive an improved bolt and 
shutter, which shall be capable of acting for a much longer 
time than can possibly be required for winding, and yet 
can be thrown out of gear as soon as the winding is done 
(which the common bolt and shutter cannot, without open- 
ing the dock-case), with a provision to secure its being so 
thrown out; and which will also have the more important 
advantage of rendering it absolutely impossible to begin 
winding without previously raising the lever to the fall 
height it is intended to go. All this can be done by a 
very simple addition to the common bolt and shutter, and 
in a manner which supersedes the necessity for any sliding 
bolt or click. 

159. The bolt B is now, as it appears in the drawings 
merely a fixed tooth on a short arm of the lever, and it is 
put into or out of gear with the great wheel, by sliding the 
arbor A backwards or forwards in its pivot holes. The 
shutter CE is no longer a cap covering the winding square 
when it is down, but a circular arc, whose centre is A« 
and which comes dose up to the winding square, so that 
when the arbor A is pulled forwards, or out of gear with 
the whed, you cannot get the key or winder on. And 


in order to prevent the lever from being merely pushed 
back, without being properly raised, its end D, or another 
arm projecting from it in any convenient place, rests, when 
the lever is out of gear, in front of the stop F (which will 
be further described presently), so that brfore you can push 
the lever back, you must lift it over the top of P, and then 
push it back, and so into gear; i^r that you can piit the 
irinder on, and as the clock goes on the lever falls, and the 
end of it descends behind the stop F, aa shewn in the side 
view of /P Q. Now F/is a kind of flap or valve turning 
on a hinge at/; and it will be seen, on looking at the side 
view of it, that you can, when you have done winding, pull 
the lever D forward, with its arbor, which will pull the flap 
forward ; and then the bolt being out of gear, you will let 
the lever drop in front of F on to the block G below it, and 
the flap immediately falls hack, so that you cannot get the 
lever back again by the way it came out, but it must go over 
the top of F, as at first, when you want to wind up again. 


The reacter may be inquiriiig, what will happen if tiie 
man omits to pull the lever out of gear : the clock mil atop 
in five or ten minutes^ as the lever will hold it fast. This 
is so done on purpose^ for it might be very easily avoided 
by making the bolt a clicks as before; but then you 
would have no security for the lever being properly thrown 
out of gear when the winding is done^ and being left in 
such a position that the man cannot begin winding the 
next time without raising it. I had originally designed it 
with a self-acting inclined plane behind F, to throw the 
lever out of gear as it descends ; but, on the whole, I am 
satisfied that it is better to omit that provision, as it would 
encourage carelessness in the man who winds; for he would 
probably never take the trouble to pull the lever out of 
gear, if he knew it would do so itself in a few minutes, 
and consequently he would always leave the clock with tiie 
double force on ; which he would take care not to do if he 
knew the clock would immediately report his negligence by 
stopping. It is obvious that this apparatus can be appHed 
as easily to any other wheel as the great wheel, though, for 
the reasons I have mentioned, I prefer that wheel. Where 
the going part winds up by a jack- wheel, as I have recom- 
mended (141) for large clocks, even when not absolutely 
necessary, the shutter C D might be put merely behind that 
whed, so as to prevent it being pushed back without first 
raising the lever; and then pulling the lever out of gear 
would, at the same time, throw out the jack-wheel, and is 
no more trouble than doing that alone, as is done in the 
apparatus which I must now describe. 


160. This mamtaiiimg power vas originally inveiitod 
for the great eqnatoreal'telescope-driving clock at Gam- 
bridgCj which lias a revolvii^ pendalum (16), and there- 
fore wiU not bear even the momentary disturbance of putting 
on and taking off an ordinary bolt and shatter, though l^e 
force is very nearly uniform when it is on. It is applied to 
the Exchange clock, and is intended by Kr. Aiiy to be ap- 
plied to the great clock at Westminster. 

The constructioii r 
of it is curious, and 
will require some 
attention to nnd^- 
stand the principle 
of its action. The 
barrel arbor C is not 
set in the clock 
&ame, bat in a 
swinging frame, of 
which one side is 
CAD, and which 

tuma on pivots of ite own (not an arbor going through 
the frame) at A, and is kept balanced between the wdg^t 
M whicb is bong to it aa shown in the drawing, and the dock 
weightWwith the weight of the barrel and ita wheels, as fol- 
lows. The pivots A are so placed that C A = the radius of 
the barrel. The reason why the sides of the frame are not 
straight will be explained afterwards : fj is the point where 


H would bang^ if the sides were straight, or where C A 
produced meets M D produced. At B is the pinion of the 
centre wheel. Now, since the dock weight W hangs hj a 
line passing through A, we may suppose the whole as a rigid 
system (there being a proper contrivance to make it do so) 
to turn round A for a short time, and we shall then have 

M Y^ as the force of M upon the pinion B ; but this is 

diminished by the weight of the* barrel and its wheels, 

which we may call N; and their centre of gravity being C, 

C A 
they will produce a force on B the other way = N -r-^. If 

therefore we suppose for convenience A ^2 to be made = A C, 

the force upon B when the system all turns round , A will 

be (M— N)t^. And we want this to be equal to the 

force exerted by the weight W when the barrel and 

. < • A C 

wheel turn round C as the fixed point, which is Wg-g: 

therefore, in order to find what the weight M must be, we 
have the equation 7^ =bc '• M=N+Wg-g. 

Suppose, for example, that N = 1 cwt., W= 2 cwt,, 
and the diameter of the barrel is half that of the wheel, 

org-Q=|' then M will have to be 4 cwt. The way in 

which the machine is all made to turn round A while wind- 
ing up, or while the weight W is taken off the clock, is 
this : there is a click K fixed to the frame C A D at A, and 
working in an internal ratchett of the great wheel, of which 
a few teeth are shown in the drawing; the common barrel 
ratchett H being fixed to the great wheel. 

The clock is wound up by a jack wheel G, working into 
a wheel set on the end of the barrel as in other large clocks; 

HB. aiet's ooing-babrel. 213 

the jack wheel arbor is placed on a level with C A, and 
therefore the point of contact A of the two wheeb may be 
considered a iBxed pointy and the whole system is still 
balanced upon it while winding. The arm AD is made 
oblique^ in order that if the equilibrium between M and W 
should be deranged by the friction of the pulleys or other 
causes, it may speedily restore itself^ a small motion of the 
point D causing a sufficiently large alteration of the length 
oi kd, or of the moment of M. Aeii^; whereas if GAD were a 
straight line, this moment would not be sufficiently changed 
until the frame had moved through a large angle, and pro- 
bably ran itself out of gear altogether. It will now be 
easily understood that this machine must be exceedingly 
expensive and troublesome to make and fix ; though the 
mere elevation or section of it in this drawing gives no idea 
of the additions and alterations which it renders necessary 
in the clock frame and the rest of the train. And after all, 
however useful it may be for a dock with a revolving pen- 
dulijun, it does no more for a dock with a vibrating pendu- 
lum, and especiaUy one with a remontoire, than a common 
bolt and shutter, which does not cost one fiftieth part as 
much as this did in the Exchange dock, provided only you 
can make sure of its being properly applied; which, as 
I have shown, may be done very easily and cheaply. 


161. The striking part of a church clock is much the 
same as that of a house dock, except in the position of the 
hanmier, and its. acting by its own weight instead by a 


spring. The shape of lu^ bells beu^ not hemispherical, bat 
of the shape in this drawing, the thick part S or sotmd bow 

of the bell is so carved, that a tangent to it wonld make an 
angle of about 35° with a horizontd plane : and therefore 
the axis C of the hammer mnst be so placed that G S will 
form that angle with the horizon. As in house clocks the 
hammer is prevented from jarring on the bell by a spring as 
shown in the drawing. 

Now it is evident that a hammer resting in this position, 
and rising in a circular arc, requires more force to raise it 
at the beginning than at the end of its motion. But as 
striking pins of the common construction must begin to act 
at some distance from the end of the lever which th^ raise, 
their force is leta at the beginning than the end of the 
motion, and so a good deal of the force of the clock is 
wasted, and the hammer is not raised nearly so high as it 
might be if the action were unifonn. As reg&<^<) the ham- 


met, we may remove this inequality in a great degree by 
putting it on a long shank^ as it will then rise a given 
height with less angular or circular motion^ and con- 
sequently with less waste of power in the train which raises 
it^ and it will also fall with less Motion on its pivots^ and 
therefore greater velocity. And as regards the striking 
pins^ we must make them of such a shape that they do not 
begin to act at some distance from^ but at the end of ^the 
lever : that is to say, they must be of the form called co^s 
or canu, of which the curve is to be determined by calcula- 
tion or experiment, so that the lever may always act only 
upon its end, and with a sliding and not a scraping friction 
between the two surfaces. 

162. Now raising a lever by cams, or co^s, or wipers, is 
a different thing from driving a wheel, though it may 
appear to be the same thing; and I may as well warn any 
reader of the common translation of 'Camus on the teeth of 
wheels,' that the rule for the construction of cams given in 
the Introduction (which is not Camus's) is entirely wrong, 
from overlooking this difference : it would be right if the 
end of the lever were a round pin. In wheels, when a tooth 
of the ' driver' has' driven a tooth of the ' follower' to a 
certain distance, they leave each other, and the motion is 
taken up by another pair of teeth ; aad when the two teeth 
part company, the end of the driving tooth is pressing 
against the side of the driven one, and not at its end, as the 
teeth G, c, in the drawing at p. 138 ; whereas a cam raising 
a lever must finish with the point of the cam at the end of 
the lever. Therefore the curve of the cam must, firstly, be 
such that it will remain acting upon the end of the lever all 


the time^ instead of begiimiiig at the end^ and then going 
£Eurther up, as in a wheels and then sliding back again ; and 
secondly^ with a view to diminishing the friction^ the curve 
must allow the end of the lever to move upon it as a 
tangent through the whole motion; and in that case^ if 
the lever wears at all^ it must still wear itself as a tangent, 
and will therefore never change its proper form, if the end 
of the lever is made a circular arc round the centre on which 
it turns. 

163. The curve which would do this accurately is called 
in mathematical language a tractrix, because of the mode in 
which (theoretically speaking) it can be described. Prac* 
tically however that method cannot be made to answer; 
neither is there any other convenient way of describing 
it, as far as I know. But it luckily happens that a curve of 
the same nature as that which is required for teeth is sufBi- 
ciently accurate for any such angle as a hammer lever has to 
move through, though it would not do for large angles."^ It 
will be found by any one who makes the calculation, that 
there will be no appreciable error, for angles up to about SO^, 
if ^ be the length of the levet or tangent, a the radius of the 
circle upon which the cams are to be set, and r the radius 

* I may remark that this curve does not raise the hammer with 
quite uniform velocity, but rather more quickly at the beginning. 
The difference however is very small, and is of no consequence ; for 
as the train always has a little run before it begins raising the ham- 
mer, it has more momentum at first. In fact, a clock made in 
the common way will often not be able to start itself, if the wheels 
are so placed that the lever is left just reating on a pin when it has 
done striking, for want of the momentum acquired by the run previ- 
ously to encountering the resistance of the hammer. 


of the cirde which is to generate the epicycloid by rolling 
on the circle of rad. a, and r is determined from the equa- 
tion, %r=^y/a*-^at—a. 

Tor example, suppose ^ = 4 inches, and as = 8, then r 
will be *9 inch, or rather less than a quarter of the length 
of the lever ; and it will not be much affected by changing 
the value of ^ to a moderate extent. These are the sizes of 
a and ^ in a large clock which Mr. Dent has lately made for 
Tavistock church, in whic^h this construction was adopted. 
There are 24 cams on the wheel of the hour part, and 36 
on each side of the great wheel for the quarters. The hour 
bell weighs 80 cwt. ; and though the wheel with these 
cams on it is of cast iron, and their surfaces are not polished 
as pins always are, the striking weight is only 2 cwt., with 
a fall of about 60 feet, which every dockmaker will recog- 
nize asi< much less than usual for eight^day clopks striking 
on much smaller bells. This arises partly from the shape 
of the Cams, and the short run of the lever upon them, and 
partly from their being on the great wheel instead of the 
second wheel, and in some degree also from the cams acting 
on the lever on the same side of its axis as the hammer 
wire, being what is called a lever of the third order; for 
in that case the pressure on the axis, which produces the 
friction, is only the difference between the pressure caused 
by the weight of the hammer, and that of the cams, instead 
of being the 9um of those pressures, as it is when the clock 
and the hammer are both pulling upwards at opposite ends 
of the lever. 

164. The striking part of the Tavistock clock; that is, 
tiie same patterns of the wheels, is adopted also in the Mean- 


213 TCEEET clocks; 

wood dock, vhich luu a bell of only 1 1 cvt. It msj be 
•Imposed that the same trheel cannot be right for a bell of 
11 cwt. and one of 30 cwt. F^haps if it had been neoes* 
■ai7 to malce a new patteni, it might haye been made of 
15 inches diuneter instead of 18, but certainly not lessj <»■ 
■ it woold have made the spaces for the cams inconvenieEtly 
■mall. And the extra size and weight are of no conse- 
qoeuce in the first wheel of the train, especially as the rope 
polls upwards. The result is that by adopting this ar- 
rangement of the cama and a cast iron wheel, one pattern 
will serve to produce wheeb £t for a bell of 11 cwt., and 
also a great deal stronger than the old-^sMoned whed 
which raiaes the hammer of great Tom of Lincoln. As I 
believe this kind of wheel is new I have given a sepaiate 
drawii^ of it, shoning a few of the cams, and the lever. 

But although the exact value of a in the above equation 
ia immaterial for the purpose of determining the shape of the 


^picycloidal cams, we have still to find out at what depth 
the cams must be placed, so that their points may just come 
up to the bottom of the teeth of the wheel, which must be the 
boundaiy of the cams. For this purpose cut out in tin or 
paper a pattern of one of the cams of greater length than 
will be required^ and also cut on the tin an edge represent- 
ing a tangent from the bottom of the cam, which will be a 
radius of the great wheel, and prolong it to a point T at 
the distance i from the bottom of the cam ; and cut out 
also on the same piece of tin an arc of the circle which 
will be described by the lever # round T (leaving a piece 
of the tin at the junction of the cam and of the arc suffi- 
dent to hold it together). Then draw on a board a circle 
about ith of an inch larger than the bounding circle and 
divide it according to the number of the cams. Take the 
tin pattern and slide it about on the board, takiag care 
always to keep the edffe which represents the radius of the 
wheel iu a rail posiL, untfl you see that a portion of 
the cam and of the lever-arc is just iucluded in one di- 
vision of the large circle. 

The place where the point T falls on the board will then 
give the exact distance of the centre of the lever arbor 
from the centre of the wheel : and another circle drawn 
on the board of the exact size of the bounding circle will 
cut the points of the cams at the proper place ; and their 
bax5ks should he cut away in circular arcs drawn with a 
radius a little longer than t. The reason why the first 
circle was drawn on the board rather larger than the actual 
bounding circle was to allow the lever to fall off one cam 

a little b^ore the next is ready to receive it. And it must 

L 2 


be remembered that the length of the lever must not be 
altered from the length t, nor the position of its centre 
altered from that determined as above ; and the face of the 
lever must be in the straight line or plane joining the 
centres of the wheel arbor and the lever arbor. If anj 
more space is required to clear the lever in its fall, it most 
be taken off the back of the cams and not off the end of 
the lever, as it is very tempting to do, or they will no 
longer work together without scraping; in fact the lever 
had better be too long than too short; and the end of the 
lever should be an arc of a circle described round the cen- 
tre of its arbor, as it will then always keep the same length 
as it wears. 

Even if epicycloidal cams are not used, neither small 
pins nor rollers should be used ; but the pins should be 
cylinders large enough for their action to begin as 
near the end of the lever as may be, and half the cy- 
linder should be cut away (as indeed every letting-off pin 
should be) to let the lever drop suddenly as soon as it 
has reached its highest point; which rollers prevent, 
causing the drdp to begin slowly, whereby part of the rise, 
and the power of the clock, is wasted. The acting face of 
the lever should be not less than half an inch broad for a 
large clock, for the same reason that broad teeth are better 
than narrow ones ; a narrow lever cuts nicks in aU the pins 
in a very short time. And in order to take as much pres- 
sure as possible off the pivots of the arbor on which the 
lever is fixed, the two arms' of the lever should be as close 
t(^ether as they can be placed ^nd not one at one end of 
the arbor and the other at the other ; and all the cranks, if 


a^y are required, should have long arms, in order to di- 
mLh the angulL motion requiri for a given ™e of th. 
hauuner ; and there should be as few cranks as possible. . 
165. I have described a simple method (20) of trying 
what force any given escapement is really using, or how 
much of the going weight is employed in merely overcoming 
the friction of the train. It may be ascertained by a 
»till more simple method, how much of the striking weight 
is lost by friction, and the motion of the train between 
every two blows, and by the hammer lever and cams bei^g 
so arranged that nearly all the power is consumed in raising 
the hammer the first inch. It appears that the hanmier 
shank, or rather the line from t}ie axis to the face of the 
hammer,' generally makes an angle of about 35^ with the 
horizon. Since we caimot help some force being lost by 
the rise taking place in a circular arc instead of a straight 
line, though the radius of the arc ought to be made as 
large as possible, we may, in comparing one clock with 
another, assume the hammer to rise in a straight line at an 
angle of 35° to the vertical. Therefore if H ia its weight 
(beyond what is required to balance the rod work, strictly 
speaking), and d the rise from the bell, the work done by 

the hammer in the day is 156 Rd cos 35° ; and if A is the 

W h . 
actual fall of the clock weight W in « days, — ia the work 

it ou^At to do in the -day if no force were lost; and the 
ratio of these two quantities will show what proportion of 
the power is lost. Cos 35°= '82, and il d is expressed in 
inches and A in feet, and n is eight days, the above ratio 

will be very nearly , which, if there were no loss of 

power, would be a ratio of equality. It wiU be found 


however that it is seldom as mnch as i^ and often as little 
as i^ in docks that strike from the second wheel. In the 
Exchange clock it is rather more, because it has cams, not 
pins. In a table given of five existing clocks among the 
parliamentary papers respecting the Westminster dock, the 
t)nly one, even if those that strike from the great wheel, 
that gives this ratio as high as | is the one in Wilton 
Place (St. Panics, Knightsbridge), which is a clock with 
east iron wheels and cams; not Uke the Tavistock whed, 
but with ten cams on each side and two hammers raised 
alternately. In the Tavistock clock the above ratio is as 
high as }; or only ^th of the power of the clock is lost in 
fidction and the necessary interval between the hammer 
Ming and beginning to rise again. 

166. It is mentioned in one of the parliamentary papers 
that in some foreign clocks the hammer is placed with its 
head downwards and its axis near the top of the bell; so 
that it is easier to raise at the beginning than at the end 
of the motion. This is no doubt an advantage when the 
hammer is raised by pins which begin to act at some dis-< 
tance from the end of the lever ; but it must be remembered 
that a hamm^ so placed will require a much larger angular 
motion than one placed as usual to raise it to the same 
vertical height, on which the force of the blow depends, and 
not merely on the distance from the bell to which it is 
raised; for it will be seen on looking back at the drawing 
of a bell that the hammer shank must stand at a mach 
larger angle with the horizon than 35** when the hammer head 
is downwards, and moreover it wiU strike the proper part 
of the bell more obliquely, since the angle of 35** (or 


thereabouts) is adopted just because it is that which enables 
the hammer to strike the bell as directly as possible. 

When the bell is not required to be swung, the hammer 
might indeed be set at the proper angle by putting the 
head upon a sort of double shank embracing the bell. But, 
as has been shown just now, we can get quite sufficient rise- 
out of a hammer set upon a shank of proper length in the 
common way, with no greater loss of power than Jth; a 
great part of which must, under any construction, be 
lost in the necessary interval between the fell of the ham- 
mer and its beginning to rise again, and in the inevitable 
loss due to the hammer spring; and therefore I think we 
may be more profitably employed in improving the construe^ 
tion of the clock itseH than in making or adopting contrive 
ances to meet a defective construction. 

. I think however that it is a question worth consider* 
ing, for a stationary bell, whether it would not be better to 
make it stand with its mouth upwards, that the hammer 
may strike it on the inside, as the clapper does. No clock- 
hammer ever gets such a sound out of a bell as the clapper 
<loes when it is ringing in fiiU swing. It is not impro* 
bable that the bell opens under a blow more freely than 
it closes : a blow on the inside makes the circle open in 
that part for its first vibration, whereas a blow on the out^ 
side is resisted by the bell as an arch. 

I may mention that a stiff hammer-spring, which allows 
the hammer to stand very near the bell, is better than a 
weaker one, which admits of larger vibrations, and therefore 
requires the hammer to be kept farther off. I have much 
increased the sound of a house cbck with an unusually 


lanre bell and hammer, by substitatinff for the common stop- 
«pSg a thin piece of v^canized InLrubber, set upon a 
firm stop just below the hammer-head. Moreover^ tlie 
hammer ought to have a broad face^ and not, as some clock- 
makers fancy, a sharp one. Besides the inferiority of the 
sound, I remember a very large and fine bell being cracked 
by a hammer set so as to strike with its edge* > 


167* When the clock strikes quarters, the striking 
wheel is made in the same way as the hour striking wheels 
only with cams or pins on both sides of the wheel, of the 
proper number for each hammer to strike 120 times in the 
twelve hours instead of 78. The cams on one side should 
not be set half way between those on the other, but the 
cam that raises the second hammer should be behind the 
other by about one-third of the distance between two 
successive cams on the same side, in order that the interval 
between the two blows of each quarter may be half that be* 
tween each successive pair of blows. In this case I sup- 
pose both the levers to ride upon one arbor, which is a bet* 
ter arrangement than putting them on different arbors, one 
below the other. In the Tavistock clock, the. two levers 
axe arranged as shown in this drawing, in order to bring 
both their long arms, into the most advants^eous position. 
I have not thought it required to show the moveable 
bushes, which are necessary to enable the arbor of the two 
levers to be passed through all the four holes in the frame 
and levers. 

In all cases, both for the hour and quarters, there ought 


to be a atrong stop (if cast on the frame all the better) for 
the long arm of the levers to strike against when the ham- 
mer falla, to take the ahock off thestriking pins, and, as far 
as possible, off the arbor. Such stops are often put to the 
short arm, and mnch t«o weak, and too near the arbor. 

The quarter bells should be the Ist and the 4th or 5th 
of a peel of 8, the hour bell being the 8th. Where there 
ate less than 8 bells, quarters may struck at every quarter 
except the 4th ; and this will only require 72 blows for 
each hammer in the twelve hours. 

168. Wherever there is a peal of ten bells, the quarters 
may consist of chimes hke the well known chimes of St, 
Mary's at Cambridge, or that (in my opinion) very inferior 
' improvement' of it at the Royal Exchange, for which, 
however, it is right to state that the maker of the clock is 
no more responsible than for the bells, about which there 
has been so much discussion. The bells used for such 
chimes are the 1st, 2nd, 8rd, and 6th of a peal of six, 
or (which is the same thing) of a peal of ten, the 10th 


feeing the hoor-beU. It cannot be done with a peal of 
eight, becanse the first six do not themsdves fonn a peal; 
and the 4th of a peal of eight is a very miserable substitute 
lor tke 6th of a peal of ten; though it is adopted in some 
places^ as at St. Clement's churchy in the Strand ; and I 
eonld mention a clock made for a nobleman a few years 
ago, who intended to have the Cambridge chimes, but the 
five bells had been cast of the notes for a peal of eight 
before the clockmaker learned (on asking me to famish 
him with the proper changes) that the bells would not 
make those chimes at alL To prevent snch mistakes in 
future, I will state what the chimes are, both of the Cam- 
bridge and Exchange quarters, and the construction of the 
barreb to produce them ; 

Cambridge* Exchange. 

^ , ( 3126 \ 1st - ( 3126 ) \ 2A 

1326 1*^*^ (l326 ( 

8d ] 6213 / 8218 J 

{ 1236 - 1st 


{The Cambridge barrel turns twice in the hour, having only 
the five changes set upon it, and the number of them to be 
played at each quarter is determined by the locking'plate, 
which turns once in the hour. The barrel must first 
be divided not into 20 but 25 equal parts, and every fifth 
division left without a pin in it, in order to aUow twice as 
much time between two successive changes as between two 
successive notes of the same change. In the Exchange 
chimes, the reader will see that each quarter begins with 
the same change, and therefore, though there are only fotir 
dianges altogether, yet the barrel must take the whole hour 

to iXifxi m> atid most have 40 pins in it^ instead of the 20 
required for the Cambridge chimes. And consequentlj, 
besides the inferiority in the tunes played by the Exchange 
dock to those of St. Mar/s^ you cannot tell what quarter 
it is until you have heard out the whole tune; whereas 
everybody in Cambridge knows directly a quarter begins 
what*it is going to be, except that the hour begins with the 
same changes as the half hour. 

Chimes of this sort are so much better than the com^ 
mon ding-dongs, and so much easier to distinguish, at least 
when they have a different tune as well as number for 
every quarter, that it is to be hoped the present plan of 
the great Westminster clock will be altered; and that as 
the hour bell is intended to be the largest ever made in 
England, viz. 14 tons, so its quarters, if they cannot be 
superior to all others in tune as well as size, will be 
at least equal to the best 'that are known. Eight bells^ 
which were proposed by Mr. Barry, are too many for 
distinctness as a quarter chime^ and so would cause great 
additional expense for no good. Mr» Whitehurst suggested 
five; but those a^re very inferior to the Cambridge four. 


169. As this subject of bells is materially connected 
with that of church clocks, I will add a few remarks upon 
it. And first, I may observe that it is sometimes a ques« 
tion whether a clock should be made to strike on the tenor 
bell of the existing peal^ or upon a little bell to be set 
upon the top of the towen When the tenor is a largd 


bell^ theare can be no doubt that it has much the best effect 
to strike on that. But then it generally requires a larger 
and more expensive clock ; and unless the tenor is a bell of 
at least 10 cwt.^ a small one of two or three cwt. outside 
the church will frequently be heard farther^ and accordingly 
that plan is sometimes advantageously adopted. 

170. Where the bells are not ready made for the 'clock 
their size is of course optional; but unfortunately their 
quality is not equally at the option of the purchasers. And 
in consequence of the disputes which have occurred respect* 
ing the Royal Exchange bells^ I wish to suggest to those 
who have to give judgment on bells^ that the tune, that is 
to say^ the note of one bell relatively to others^ is a totally 
distinct thing from the tone, or absolute quality of the bell^ 
and of infinitely less consequence; because the note can 
easily be altered sufficiently^ but the tone of a bad bell can 
never be mended^ except by breaking it in pieces^ and melt- 
ing it; and not always by that^ if the metal has been bad 
originally, or the bell-founder does not know how to make a 
good bell. In order to judge of the absolute goodness or 
tone of a bell, what is wanted is not so much musical know- 
ledge or perception of tune as experimental knowledge of 
what bell-metal is capable of. A peal of cast iron bells 
might be made perfectly in tune, and to a person who had 
never heard bell-metal bells, would appear a perfect peal. 
No rules can be given to enable people to judge of the 
quality of sounds ; but a few things may be mentioned as 
necessary to attend to; such as, whether the bell sounds 
freely on being hghHj touched; how long it 'holds the 
sound,' compared with other known bells of about the same 

BELLS. 229 

«2e, and of good quality j and particularly whether on filing 
or polishing the bell anywhere the metal appears perfectly 
dose and free from holes. If it does not^ you may be sure 
the bell is a bad one^ without any further examination^ and 
it ought to be condemned at once. A bell may also be too 
thin for its size^ and perhaps occasionally they are made 
too thick. In bad bells^ it may be frequently observed that 
you hear a harsh metallic sound of the knock of the ham- 
mer, independently of the continuous or ringing sound, 
which alone ought to be heard. Other points must be left 
to the discretion of the judge. 

As many persons may have read some remarks on the 
relative merits of the great bells of St. Paul's and Christ 
Church, Oxford, in one of the ParUamentary papers, I must 
say that I do not agree with them; on the contrary, I 
think St. Paul's far the best of the four large bells of Eng- 
land, though it is the smallest of them, being about 5 tons; 
while York is 12, Lincoln 6i, and Oxford 7 J, which last 
is a remarkably bad bell. There is a general opinion that 
bells cannot be made now equal to the old ones. It is true 
that bells improve in sound for a few months, but no more; 
«nd they only alter in loudness, not in the quality of the 
tone. The badness of many modem bells is due not to 
want of age, but to want of skill or attention in the founder. 
I have seen as bad bells as need be, with dates of about 200 
years ago ; and the best I ever heard (for a small peal), at 
Castle Camps, in Cambridgeshire, were made about 20 
years ago by a country bell-founder, of the name of Dob- 
son, who however did not meet with sufficient encourage^' 
ment^ and lately died a pensioner in the Charter House; 


and many exoellent beUs^ both large and small, Were east 
by Mr. Mears, and his predecessors^ Messrs. Leicester^ 
Pack, and Chapman, at the weU-known foundry in White* 
chapd, which for some years has enjoyed nearly a mono* 
poly ; I have lately however seen some very good bells^ 
made by Messrs. Taylor of Loughborough ; and I have had 
the opportonity of comparing one of their beUs with a 
foreign one imported by Mr. Dent for trial, and the Eng« 
lish one was decidedly the best. And therefore, though 
the casting of bell-metal is a 'mystery^ requiring consider- 
able skill and management beyond merely melting together 
certain quantities of copper and tin> there is no reason 
to reckon bell-founding yet among the artes perditce* If 
people would always reject bad bells, good ones would soon 
become as common as they ever were. 


171. It is usual to make the quarters let off the hour ; 
that is, the locking-plate of the quarters is furnished with 
a pin or a snail which, while the last quarter is striking, 
performs the office of the discharging pin on the hour wheel 
•of the going part when there are no quarters. And for 
ordinary clocks this plan does well enough ; but it is evi* 
dent that in that case you cannot rely on the first blow of 
the hour, which is the proper indication of the exact time, 
being right to several seconds, because it depends on the 
rapidity with which the quarters may happen to strike, in 
addition to the ordinary sources of inaccuracy in the time 
which 4he &our train takes to get into action; both of which 


differ considearably in different states of the clock. And 
therefore in some clocks there are two snails set on the 
honr'^wheel^ one of which lets off the quarters^ from a quar- 
ter to half a minute before the hour^ and the other snail 
lets off the hour striking part exactly at the hour. And 
here I may remark^ that the larger the snail is the more 
accurately it will let off^ because the linear space on its 
circumference. corresponding to one beat of the clock will 
be larger^ but the more friction it will cause. Moreover^ 
if the time of striking the first blow is intended to be very 
accurate^ the hammer should be left on the liffc^ or just on 
the point of fallings instead of having to wait while the 
train is raising it. And in that case (and indeed it is better 
to do so whether the hammer is left on the lift or not) 
there should be a small click applied to one of the wheels 
of the striking train to prevent it goiag backwards when 
the clock is being wound up. Mr. Dent uses merely a 
single pin^ set as a tooth into the fi^y arbor^ and so placed 
that the click falls against it when the clock has done 


172. This naturally leads to the subject of remontoireB 
in the train, which I have several times referred to> and 
which are quite distinct from remontoire escapements. If 
the scape-wheel^ instead of being driven directly by the 
train and the clock weighty is driven by a small weight 
which is wound up at every twenty or thirty seconds by the 
clock train, the clock will have these advantages : lat. The 


scape- wheel will be driyen by a force as nearly uniform as 
possible^ being free from all the inequalities of the friction 
of the train and dial work^ and the effect of the wind upon 
the hands. 2ndly, The striking may be let off more ex- 
actly than in the common way^ because the snail will turn 
through a fifteen times larger angle if let off by the re- 
montoire every thirty seconds than by the pendulum every 
two seconds ; and so you may be quite sure that the striking 
will be let off at the sixtieth second of the sixtieth minute 
of the hour, which you cannot secure in the common way. 
Srdly, The long hand will move a visible distance by jumps 
at every half minute, which will enable a spectator to ob- 
serve the time as accurately as from the second-hand of a 
regulator ; whereas in common turret clocks the time can 
hardly be taken to less than a half a minute even in the 
most favorable positions of the minute-hand. In short by 
the use of a remontoire in the train a dock would go 
better, and would also be available as a regulator both by 
sound and by sight, as perfectly as an astronomical clock ; 
while a common turret clock, if it goes ever so well, is 
useless for very exact observations, such as a person wanting 
to regulate a good clock of his own would require, as he 
can neith^ tell the time from the hands nor rely on the 
striking to several seconds. 

178. There have been various contrivances fortius pur- 
pose. One is described by £eid as having been put up by 
him at Edinburgh in the last century, which was worked 
by a Huygens's endless chain (79). He says it went very 
well, but that the chain and other parts connected with it 
wore out so fast that it was removed. There have been 


many attempts at it in France^ but they are stated in Erench 
books on the subject to have been irenerally unsuccessful. 
However a public clock with a remontoire i the train ha. 
now been going for five or six years in London; and the 
efficacy of it as regards time keeping may be judged of by 
any of the numerous chronometer-makers who live within 
sound of the Exchange bells; and any person may judge 
of the effect of it as regards the other objects I have men- 
tioned by looking at the long hand just before it strikes the 
hour : he will see the last jump of the hand, after the quar- 
ters have done striking, take place at the same moment as 
he hears the first blow of the hour. And as the great 
clock for the new palace at Westminster is to have some 
contrivance of this kind, a description of it will probably 
be interesting. 

Any person with a moderate amount .of ingenuity will 
.see.that the only real difficulty iji]the problemis to keep. the 
action of the remptitoire weight on the scape-wheel while it 
is being lifted. Harrison's going ratchett might of course 
be applied; and although it would cause a small amount of 
friction, it would have some advantages perhaps over the 
one I am going to describe. The plan of the Exchange 
dock remontoire is this, omitting some merely mechanical 
details. The wheel D, whose pinion is driven by the centre 
wheel £, has int^iud teeth; and a wheel €, which rides 
upon the arbor of D, has external teeth as usual. On the 
arbor of D there rides also ah arm or lever G B A, carrying 
the remontoire weight A; and there is (or may be con- 
sidered to be) a pin or stud upon the arm for a wheel B to 
lide upon, which works between the ext^nal teeth of G and 


t^e uLternal teeUi of D. Consequently, whenerer the wheel 
D ia driven by the train it will raise GiB right aie of B, 
which will raise the lerer and the weight; bat t^e left side 
of B will nererthdeas be always pressing downwards on C 
and tending to tnm it the opposite way to D ; and therefore 
a large wheel F fixed to the whed C may be employed to 
drive the scape-wheel^ which will then be driven merely 1^ 
the remontoite weight and not by the clock train. 

174. The mode of letting off the train at every twenty 
seconds is this. The scape-wheel arbor tnms in a minute 
and is nearly half cut throogh in three places near together; 
and on the rim of B, which is made broad enough for the 
purpose, are placed three sets of spikes in diffa^nt planes 
corresponding to the three nicks in the scape-wheel arbor; 
and these three nicks being made at uiglea of 60° to eadi 


other, they come successively at intervals of twenty seconds 
into such a position that a spike of one set can pass through 
the corresponding nick, hut a spike of the next set strikes 
the arhor in a place where it cannot get through for 
twenty seconds more, and then a spike in the third plane i$ 
stopped in the same way, and so on. Thus at every twenty 
seconds the wheel D is allowed to move through the space 
between two spikes, and so the Uttle weight A is raised 
through a small space by the force of the clock train, and 
the forcQ on the scape-wheel is the same upon the whole 
for every successive third part of a minute, though during 
the twenty seconds it varies a Uttle as the distance of A 
firom a vertical line through C varies, by reason of its des- 
cribing a circular arc. 

Half a minute would be a better interval than twenty 
seconds, because it is not easy to see at a distance whether 
the hand is ten seconds before or after the half minute^ 
and the motion at each move would also be half as much 
i^aiii for thirty seconds as for twenly. 

175. The plan which is adopted in the French turret 
docks, which I understand are now generally made with a 
remontoire in the train, is the same in principle, only in- 
stead of a wheel running between the internal teeth of one 
wheel and the external teeth of another, a bevelled wheel 
riding on the remontoire arm runs between two other 
wheels at right angles to it, as shown in the next drawing. 
The wheel D performs the o^ce of the internal wheel D in 
the Exchange clock, raising the weight A by means of 
the wheel B, which turns freely on the remontoire arm, 
and so always presses dQwnwards on C, to which is fixed^ 


tubrbt clocks. 

as before, ihe 
wheel P irhich 
is to drive the 
scape - vheel ; 
and the broad 
wheel, with the 
three rows of 
spikes in it, is 
fixed to D, as 
before. I do 
not know that 
eith^ of these 
methods is bet- 
ter than the other, except that, as it is impossible to cat 
bevelled teeth to the right shape as traly as flat ones, there 
is probably rather less Motion in the internal wheel plan 
than in the other ; but it is also the more expensive of the 

But there is this objection to both of than, that the 
■pikes strike the acape-wheel arbor with considerable force, 
and even while they are at rest press upon it pretty heavily, 
and therefore a larger force is required to drive the scape- 
wheel. The force of the blow has been somewhat dimi- 
nished in the Exchai^ clock by putting a spring with a 
concave face for the spikes to slide up before they reach 
the ubor. In tiie more recent French clocks, it is done 
by making the train drive a fly, which moderates its velo- 
city ; tuid iu some of them the spikes are pnt on the end 
of the fly, or an arm on the fly arbor, which is a very great 
improvement, as it also diminishes the constant pressure 


and friction on the arbor of the scape-wheel^ in the ratio of 
the velocity of the spikes when set on the second wheel to 
their velocity when set on the fly. They have also been 
■^. .0 ... Iff by . Wer lie . .^g pL, inete^i o, b, 
nicks in the arbor; but that method gives the scape-wheel 
more to do^ and produces much greater friction. This 
IjLOwever would not signify where there is a remontoire 
escapement as well as a remontoire in the train ; and per- 
lliaps for a very large clock with such an escapement this 
might be found the best way of letting off the train remon- 
toire^ since in that case the pressure or blow of the spikes 
is immaterial^ as it does not fall on the scape-wheel arbor. 


176. A very ingenious and curious application of the 
gravity remontoire has lately been made by a French clock- 
maker^ named Wagner^ for the purpose of obtaining a con- 
tinuous motion of the heavy part of the train combined with 
the accuracy of a vibrating pendulum driven by a constant 
force. Instead of the remontoire being let off at definite 
intervals by the scape-wheel, the bevelled wheel D works a 
fly, by the intervention of two or three intermediate wheels. 
This fly revolves horizontally below the clock frame, for a 
reason to be explained presently. If the fly always turned 
at a constant and proper rate, it would evidently let the 
driving wheel D turn one way just as much as the opposite 
wheel C turns the other way, under the action of the escape- 
ment and the remontoire weight, in any given number of 
seconds; or, in other words, those two wheels between 



them would always keep the remontoire arm at the same 
height. Now this arm is prolonged to a convenient length, 
and from its end is hung^ by two wires, a thing like a gonff, 
suspended horizontally, and with a hole in the middle for 
the fly arbor to pass through; and this gong, or bell, is so 
placed, that when the remontoire is at its medium height 
the beU. about half covers the fly, which tarns within it : 
if the remontoire falls below its medium height the bell 
will evidently fall, and more completely enclose the fly ; 
and the effect of this is that the air within the beU becomes 
rarer by the action of the fly, and consequently offers less 
resistance to the fly, which will therefore regain its proper 
velocity ; and of course an increased velocity of the fly is 
identical with an increased velocity of the traiu, which was 
previously going too slowly to keep up with the regular 
motion of the scape-wheel. And in like manner, if the 
train and fly are going too quickly, the bell is lifted up, 
and the fly more exposed to the external air, which offers 
an increased resistance, and so again restrains the velocity. 
There is indeed, theoretically, a much more simple way 
of combining a continuous motion of the train with a 
vibra-ting pendulum. Por if any point in a heavy pendulum 
is connected by a long horizontal rod with a crank, or a pin 
set on the face of a wheel, moving in the same plane as the 
pendulum, and the pendulum is made to swing just so fiEir 
as to let the crank turn round iu one double oscillation, it 
may be easily proved that the natural velocity of the given, 
poiut in the pendulum is the same in every position as the 
velocity, iu a horizontal direction, of the end of the crank, 
the crank itself revolving uniformly. The practical diffi- 


cnlty^ however^ is to keep the pendulum always vibrating 
the same arc^ when there is a variable force acting on the 
crank, without which it would not answer. 


177. Instead of a weight acting on one of the wheels of 
the train, as in any of the preceding methods, a spring may 
be used to communicate the force from the pinion of the 
scape-wheel, or the wheel below it, to the wheel itself. 
And a spring possesses these great advantages over a^ 
weight, that it requires no maintaining power to keep it in 
action when winding up, as I explained in the case of a 
spring clock without a fusee (80), and that it acts without 
any friction. These advantages are so great, that I have 
no doubt a spring remontoire may be made better, and 
a ^, ^j J^ eh»p», 1. », g.^ ^ 

There is a description of a spiral spring remontoire let 
off by nicks and spikes, as before described, in the Ency* 
clopsedia Britannica. And I have lately seen some small 
Prench clocks with a spring remontoire on the second 
wheel, and I am told at Mr. Denf s they go better than 
ordinary French clocks without a fusee, which, however, is 
not saying much for them. The spring should, if possible, 
be put on the arbor of the scape- wheel instead of the 
second wheel, because then it acts without any friction 
of the scape-wheel pinion ; in fact, the scape-wheel then 
has no pinion, properly speaking. 

But all the remontoires of this kind that I have seen or 
read descriptions of, are liable to the objection that either 
the wheel or its pinion rides upon the arbor of the other of 


them, as in Mr. Air/s eacapement (47), which caxiBes the 
wheel to turn with considerable friction ; and from this and 
other causes, the spring remontoiies that have been hitherto 
made, especiaUy in turret clocks, do not appear to have 
given satisfaetion. I shall therefore describe one which has 
been lately made, and in which the objection I have just 
now mentioned is removed, and some other advantages 
obtained, and which is perfectly easy to constmct, aoid at a 
moderate expense. 

The lai^ wheel in this drawing, which is the second or 
centre wheel (there being only three in the train} drives a 
pinion of 14 leaves, which ridea on a stud or fixed arbor 
screwed into the clock frame; and the pinion is also pro- 
longed into a pipe large enough to enclose, without toneh- 
ing, the brass bush, which is fixed on the end of the stud 
to form a pivot-hole for the scape-wheel arbor. On this 
pipe is fixed a spiral spring, (the mainspring of a moderate- 
sized musical box), of which the outer end takes hold of a 


pin set on the back of the scape-wheel. It is evident there- 
fore, that if the pinion and pipe are turned round or wound 
up from time to time, it will wind up the spring, and the 
sca^e-wheel wiU be driven by the spring without any fric- 
tion at all, except that of its own pivots ; and also that 
whenever the second wheel is let go, it will so turn the 
pinion and wind up the spring. 

The letting oiBF is not done in the same way as in the 
Exchange clock, but the wheel drives another pinion 
(marked 7) on an arbor which carries a fly; and this fly, 
(though not so exhibited in the drawing) may have at its 
ends the spikes which axe to slip through the notches in 
the scape-wheel arbor as before described. Now, as- 
suming the scape- wheel to turn in two minutes (which is 
more convenient than one miuute for turret clocks) and the 
lemontoire to be let off every half minute, the letting off 
would be done by a four-armed fly set on a pinion of the 
same number as the scape-wheel pinion, and four notches in 
the scape- wheel arbor acting as the three do in the Exchange 
clock. But I have also made an alteration in this respect, 
for the purpose of diminishing still further the pressure of 
the spikes on the scape- wheel arbor, and getting a larger 
motion of the fly in proportion to the same motion of the 
train, and consequently a slower motion of the train during 
its action. The scape-wheel arbor comes through the frame 
and ends in a cylinder, of which the face has two nicks cut 
across it, one broad and i inch deep, and the other narrow 
and i inch deep; and therefore if one end of the fly 
has a short and broad pin properly placed, it can slip 
through the broad nick only ; and a long and thin pin 



at the other end of the fly will slip through the narrow 
nick only, when the nicks respectively come into a posi- 
tion at right angles to the fly. In this way the two- 
armed fly is allowed to tnm half round at every quarter 
of a revolution of the scape-wheel ; and as the fly pinion 
has seven pins while the other has fourteen, the remon- 
toire spring is wound a quarter round at every l^t off of 
the remontoire. The second wheel, which drives them 
both, turns in eighteen minutes, and consequently if the 
spikes were set on it instead of on the fly, their pres- 
sure on the scape-wheel arbor would be 36 times as great. 
I found this pressure amount to barely two ounces, with 
a very heavy weight on the clock ; and as it is made to act 
upwards, it only relieves the pressure of the arbor and the 
cylinder on the adjacent pivot-hole. 

The spikes or pins are not really set on the end of the 
fly, but on bent springs about three inches long, in order 
to diminish the force of the blow on the scape-wheel arbor, 
and the too sudden stoppage of the train ; and for the pur- 
pose of doing the same thing more completely, and also 
stopping the recoil of the fly after the blow, there is a fric- 
tion spring set on the frame, which the fly has to slide over 
just before it reaches the cylinder; this of course diminishes 
its velocity or momentum when it is greatest, and also acts 
as a click which the fly cannot pass in recoihng. 

It remains to be explained how the adjustment of the 
remontoire spring is made, in case it is found to give too 
much or too little force to the scape-wheel. The fly is not 
fixed to its arbor, but it has a click that takes hold of a 
ratchett with ten or twelve teeth, fixed to the arbor in the 


usual manner of the striking fly ratchett. Therefore when 
you want to alter the tension of the spring you have only 
to lift the click, without stopping the clock, and shift the 
ratchett as many divisions as may be necessary, which will 
turn the scape-wheel pinion and spring just half as much 
as you turn the fly arbor, since the scape-wheel pinion has 
twice as many pins as the fly pinion. The ratchett is made 
with square teeth, which are also numbered, for greater 
safety of action and certainty in adjusting it. The spring 
must of course be occasionally cleaned and oiled, to keep 
it from rusting, and so should the pendulum spring. 

This is the construction of the clock represented in the 
;-&'ontispiece, which is just completed for the newly built 
church at Meanwood, near Leeds. It is of course too soon 
as yet to give any account of its performance as regards 
time-keeping, but its mechanical action is quite satisfactory ; 
and as that is the only novelty in it, the principle of the 
spring remontoire having been tried for some time, I see 
no reason to doubt that it will be altogether successfuL 
The going weight required is about half as much again 
as it would be if there were no remontoire, because all train 
remontoires require additional weight to overcome the ad- 
ditional friction, and to wind up the remontoire with- 
out any hesitation; but as none of this additional force 
reaches the escapement it is of no consequence, and you 
can put on weight enough to drive the hands in all weathers 
without at all affecting the swing of the pendulum; whereas 
in a clock without a remontoire you cannot put on weight 
enough to drive the hands in bad weather and when the oil 

is frozen, without making the pendulum swing too far as 



soon as it becomes wanner. I have seen the semi-arc of 
the pendulum of a church clock, and a very good one, vary 
nearly a degree between winter and summer. 

Another effect of this remontoire is the remarkable silence 
of the beat, as compared with the Tavistock clock in which 
the wheels are of the same size, or with the Exchange clock 
in which the momentum of several wheels and the re- 
montoire weight has to be stopped by the pallets at every 
beat ; whereas in this clock the only thing to be stopped is 
the scape-wheel, which is less than four inches in diameter 
and only moves 4^° at each beat. It is probable also that 
it follows the pallets more closely when moved by a spring 
than when, moved by gravity. This silence of course in- 
dicates that there is less waste of force at each beat in a 
clock with a spring remontoire of this kind than either 
without any remontoire or with a gravity remontoire. 

178. I have hitherto spoken of the ^(?fo^t<?a^ advantages 
of a train remontoire. But it has besides the economical 
advantage of superseding the necessity for highly finished 
wheels in the lower part of the train, as it reduces that 
part merely to a machine for driving the weight of the 
hands and winding up the remontoire, instead of being a 
machine for transmitting to the pendulum as uniform a 
force as possible. And therefore I see no reason why the 
great wheel, hour-wheel, and all the leading-off or bevelled 
wheels (of which sometimes as many as nine are necessary) 
should not be of cast iron instead of brass, which it seems 
would almost cause their cost to be measured by shillings 
instead of pounds. If it had been certain beforehand 
that this new remontoire would answer, the Meanwood 



clock would have been so made. I know that there is a 
prejudice against cast iron clock wheels even in the striking 
part. But people would easily see that this prejudice is 
unfounded if they would only remember that the striking 
part of a clock is nothing but a machine for raising a heavy 
hammer 156 times in a day, or rather in ten minutes, since 
it would strike 156 in about that time if allowed to go on ; 
that all you want is sufficient force to do this without any 
unnecessary wear of materials ; that (as I have shown) cast 
iron clocks are now doing this work with considerably, fe** 
waste of power and therefore less wear of materials than 
the best brass clock probably in the world ; and lastly that 
such clocks can be made much cheaper than brass ones 
equally good. And when the going part is also reduced, 
as it is by a remontoire, to a mere weight-raising machine 
in which uniformity of force is of no consequence, there 
will be no more necessity for brass wheels there than in the 
striking part. The only question was whether a remon- 
toire could itseK be made at a moderate expense ; and it 
now appears that it can ; indeed I was led to turn my at- 
tention to the contrivance of such a remontoire by a 
remark of Mr. Denfs, that a good and cheap remontoire 
was essential to any further material improvement in turret 
clocks, especially large ones ; and it may be added that a 
good (that is, a secure) and cheap maintaining power or 
' going barrer apparatus was equally essential to the making 
of such large clocks except at an enormous increase of their 
cost, and I think that difficulty has also been removed as 
described in § 159. 

179. I have spoken throughout of bra^s wheels; 


meaning thereby either brass or gun-metal, which is a com- 
pound of copper and tin instead of copper and zinc^ with 
a lower proportion of tin than in bell-metal. The pro- 
portions for the best gun-metal are^ I understand^ &om 
92 copper -I- 8 tin to 90 copper -f- 10 tin; and such metal is 
harder than the best hammered brass^ and yet not too 
brittle for the teeth of wheels. It is however a difficult 
metal both to make and to work : the castings^ especially 
large ones^ being often full of defects, consisting not only 
of holes, but of hard pieces which are as destructive to 
wheel-cutters as a cinder in a piece of bread is to human 
teeth ; and it is in any state a very tough metal and clogs 
the files. I am informed however by the best authority on 
this subject, that these defects in the casting may be pre- 
vented by not putting the tin to the copper at once in the 
above j^oportions, but previously making an alloy of 2 
copper + 1 tin, which they call 'hard metal,' and which is 
the ' highest' compound the metals admit of without the 
excess of tin beginning to make the alloy softer again : to 
this alloy the further proportions of copper are afterwards 
added; and in order to obtain the greatest density or 
specific gravity which the compound admits of, the castings 
are always made with a long ' dead-head' to produce con- 
siderable pressure on the fluid metal. I am told also that 
adding a very small quantity of zinc, about 1 per cent., 
makes the metal both more fluid in the crucible and more 
compact when cold. The Specific gravity of metal cast in 
this way at Woolwich has been known to be as high as 
9, and is generally about 8*75, while that of brass is only 
8*37 (according to the ' data' in Weale's Dictionary). 


180. I think it is also worth the consideration of clock- 
makers whether they could not get more true and sound 
cast^igs^ either of brass or gun-metal^ and such as would 
suffer less waste and require less labour in squaring out and 
turning up, by having cast-iron moulds made for the sizes 
of wheels they are in the habit of using. Large castings, 
even in brass, are often defective for as much as ■^^th of an 
inch deep ; and I understand a very sharp and clear cast- 
ing can be obtained from a heated iron mould instead of a 
damp sand one, since it does not chiU the metal too soon 
but lets it run into the smallest comers. The part of the 
mould between the spokes would have to be made separate 
and bevelled a little, so as to lift up, and let the wheel 
contract in cooling, or it would probably break itself; and 
the mould should not be an open one, but close, and with 
a longish pipe to it, by way of a ' dead head,' to produce 
pressure upon the metal, as the effective part of bells, viz., 
the mouth, lies at the bottom of the mould. These how- 
ever are only suggestions, which some person better versed 
in the art of brassfounding may perhaps reduce into a more 
practical shape. The necessity of some improvement in 
this art will be evident when I state that three brass cast- 
ings were made for the great striking wheel of the Tavistock 
dock before one was obtained sound enough even to cut 
for the model ; and more, I believe, for the great wheels of 
the Exchange clock, which are of gun-metal, as those of 
the Parliamentary clock are required to be, in my opiaion 
very unnecessarily as regards all but the escapement part of 
it, in such a large clock : I mean of course that they should 
be of iron. This diflBculty of getting really good gun-metal 


castings^ hard enoagh to be any better than brass^ causes 
some, clockmakers to prefer, well-^hammered brass to such 
gun-metal as they can be sure of obtaining from the foun- 
ders. At any rate however the bushes, which are too 
.thick hardened throughout by hammering, and in 
wliich brittleness. does not signify, ought to be made of 
guurmetal; but it should be free from hard bits, whidi 
would probably cut the arbors. It must not be forgotten, 
moreover, that gun-metal is more expensive than brass, both 
in the making and the working. 

181. Lastly, all the iron work> whether cast or wrought, 
and whether visible or invisible, except just the acting sur- 
faces, ought to be painted. Polishing it does no good, cbstis 
money, and requires continual cleaning and oilmg to keep 
it from rusting. This rule has been followed throughput 
in the Meanwood clock, though it is one in which, as the 
reader has seen, no expense was spared that was thought 
likely to produce any useful effect. 

And with regard to this rather important matter rf 
oiling, fresh oil ought never to be put on, especially on the 
escapement, without thoroughly wiping the old oil off. I 
have had to take out a scape-wheel and wash it with soda to 
get off the accumulations of oil with which an attentive 
sexton had lubricated it in the course of two or three years. 
The pivots should have clean oil whenever they appear to 
be dry; but. the pinions should only be wiped with an oily 
cloth, and not have: oil dropped on to them, as is the com- 
mon practice; and this applies to house.clocks as well as 
church clocks. Asfaras my experience goes; I think no 
oil equal upon the whole to some which was recommended 

OIL. 249 

to me at Mr. Dent's for the purpose of preserving steel 
surfaces from rusting: it is neafs foot oil, well-stirred 
about in water, skimmed off and filtered once or more 
through blotting paper until it is quite clear. Though it 
is the most filthy looking stuff at first, it produces at last 
an oil more fluid than the best sweet oil, and which has 
the important advantage of never freezing, nor forming a 
brown cake on iron, as the vegetable oils do; I should 
think, therefore, it is peculiarly suitable for church clocks : 
indeed I have applied it to a house regulator for some 
time, and, as far as I can judge, with very good effect. 

18ii. The pulleys for church clocks require more atten- 
tion than they frequently receive, especially as the fixed 
pulleys are often put high up in the tower, where the 
oiling of them, is neglected. Every pulley ought to be as 
large in diameter as it conveniently can be made, for the 
obvious reason, that the larger it is the slower it has to 
turn. I should say that no pulley for a moderate sized 
clock ought to be less than nine inches diameter, nor less 
than twelve inches for a large clock with heavy weights. 
Another point, which is of perhaps even more importance, 
is, that the pivots, or arbor, should be fixed into the 
pulley, and not a pin put through a hole in the pulley, as 
the common pulleys for raising weights are made, in which, 
for various reasons that I need not specify, it is of less 
consequence than in clocks. For when the pin goes 
through a hole in the pulley, it requires to be much thicker 
than pivots turning in bushes set in the frame, in order to 
obtain the same strength; and therefore the friction is 

greater, and the weight required to drive the clock greater, 



and the wear of the pin and the hole in the pulley greater 
than it need be. Secondly, the effects of this wearing are 
much more mischievous ; for pivots always pressing in one 
direction only wear a hole larger in one direction, or make 
it longer, and it will therefore work in it with no more 
friction at last than at first; whereas when the hole in a 
riding pulley is worn bigger, the friction is increased as 
much as if the pin had been originally made so much 
larger; and indeed more, because the bearing surface of 
the pin is narrower than if it were as large as the hole. 
Lastly, a fixed arbor, with brass bushes screwed into the 
pulley frame, keeps the pulley much steadier, and with 
more equal bearing on both sides, than a pin. Pnlleys 
made as I have described of course cost rather more than 
those made in the common way, but they are very well 
worth the additional expense. 

183. There are two or three other matters which should 
be attended to in putting up church clocks. One is to 
enclose them in a wooden box that locks up, leaving the 
winding squares open, so that the clock can be wound with- 
out unlocking the case, and so that nobody but the person 
who has the care of the clock can get at the works. The 
pendulum should also be enclosed all the way up. And it 
is essential to the keeping of a clock in good order that 
the place where it stands should be light : if it is not 
light already, it should have a window put in before the 
clock is put up ; and if the architecture will not allow it 
anywhere else, the middle of the dial may be, and often is 
with good effect, made a window. When the clock is a 
good one, worth looking at, it is a good plan to have some 


smiall panes of plate glass put in the doors of the box or 
case, so as to show the escapement and the regnlating dial, 
which, if the clock stands in the belfry, will serve also for 
a dial for the ringers. If the bottom of the pendulum 
comes in a convenient place, it is as well to put a piece of 
glass to show the degree plate, so that anybody can see 
how far the pendulum is swinging. All these little things 
tend to make people careful of their clock when they have 
got a good one. Mr. VuUiamy mentions several public 
clocks in Paris which are fitted up with glass, so that they 
can be completely seen without opening them; and the 
same thing has lately been done at the Exchange clock, 
and is contemplated by the astronomer royal for the Par- 
liamentary clock. 

184. Not long ago, I heard that the weight of a church 
clock in London had broken the rope and fallen (as a 
weight of several cwt. easily would do) through the belfry 
floor and the pavement of the church into the vaults below. 
This accident, which is not very unconmion, is most likely 
to arise from the sudden stoppage of the weight when it is 
wound up until the sexton feels he can wind no farther ; 
and with the view of stopping him in time, a string may be 
put to each weight and over a small pulley at the top, and 
the other end, with a little weight attached to it, made 
to come down close by the place where the clock is wound 
up ; so that when the great weight is up, the little weight is 
down, and the man will know that it is time to stop wind- 
ing. This is like the weight put to an organ to show the 
blower when the bellows are fall. Where the weights come 
down on to a floor over the heads of people in the church, 


it is prudent to put a large box filled with wool or sawdust 
on the floor, where they would fall; the wool or sawdust 
need not be left open, but should be covered up with 
boards, which will only break first if a weight falls. 



185. I have no doubt that it will be interesting both to 
clockmakers and to other persons to know what has taken 
place up to this time with regard to what may be called 
the National Clock, both on account of the building on 
which it is to be placed, and because it is intended to be 
the best and largest public clock in England, and in the 

Six years ago, viz. in March, 1844, Mr. Barry, the 
architect of the new Houses of Parliament, wrote* to Mr. 
Vulliamy, to ask if he would furnish him with a plan for the 
clock; and as he could not undertake that Mr. Vulhamy 
should be employed to make it, he inquired upon what 
terms he would furnish the plans, first, in the event of his 
being employed to make the clock, and secondly, of his not 
being employed. Mr. Vulliamy, in reply, proposed that, if 
employed, he should be paid 100 guineas for the specifica- 
tion, calculations, working and other drawings ; and if not 
employed, an additional 100 guineas for his time and 

* I may as well state that I shall follow the words of the corre- 
spondence as nearly as I can conveniently, and without perplexing 
the reader, printer, and myself, by contyiual quotation marks, and 
interruptions of them. 


' Shortly af(:erwards Mr. Barry wrote to the Board of 
Woods and Forests, saying that, owing to the progress 
making with the clock-tower, it was desirable to have the 
necessary specifications, working drawings, and estimates 
prepared ; and that he had therefore applied to Mr. Vulliamy 
as, in his opinion, the person best qualified to make out such 
specification, working drawings, and estimates ; and he for- 
warded the two preceding letters, and recommended that 
Mr. Vulliam/s offer be accepted. In a few days he re- 
ceived an answer from the Board conveying to him the 
requisite authority to engage Mr. Vulliamy upon the service, 
and upon the terms named in Mr. Barry* s letter. This was 
of course communicated to Mr. V., who replied that he 
should proceed without delay to prepare the plan of the 
clock for Mr. Barr/s inspection. 

In January, 1845, however, Mr. Vulliamy wrote to 
Mr. Barry, to say that he had just observed a mistake 
\ in the letter from the Office of Woods and Forests, viz., 
^ that though Mr. Y. had said nothing about an estimate, 
\ the Board had incidentally introduced the word, no doubt 
I as a matter of course ; and that to make an estiniate would, 
^ until it is definitely settled how the clock is to be made, be 
quite useless; and would be a work of great labour, occu- 
pying much time, and not contemplated either by himself 
or Mr. Barry ; and that he thought it right without delay 
to notice the circumstance, lest it should give rise to any 
misunderstanding afterwards. 

To this letter there appears to have been no answer from 

Mr. Barry, or notice of it by the Board, except so far as the 

\ subsequent letter of Mr. Barry, of July, 1846, is an answer. 


In November, 1845, Mr. Dent wrote to the Board, to 
say that he was desirous of being admitted as a candidate 
for supplying the large clock, and such others as might be 
required for the new Houses of Parliament ; and by way of 
recommendation he referred among other things to the tes- 
timonials of the astronomer royal and ,Mr. G. Bennie, re- 
specting the Exchange clock, and proposed to obtain the 
sanction of the Board for erecting the new clock, subject to 
the approbation of the astronomer royal, with Mr. Barry, 
and Sir John or Mr. George Eennie, as referees. The 
Commissioners answered that when the drawings and sped- 
fications for constructing the great clock were completed 
and could be submitted to the several clockmakers who 
might be applied to, as the basis upon which their tenders 
were to be founded, the Board would include him as one of 
the competitors, for making that as well as any other clocks 
that may be subndtted to competition. 

To this Mr. Dent replied, that if adherence to draw- 
ings and specifications to be prepared by another clockmaker 
were to be stringent on him, he must decline to become a 
candidate; that he should feel it a duty to comply with any 
suggestions from the astronomer royal, but could not 
engage to act under the directions of authority less emi- 
nent, or to follow instructions, which by degrading him to 
the position of a mere executive mechanic, would prove de- 
trimental to his reputation. 

Apparently in consequence of this letter, though after 
an interval of some months. Lord Canning wrote to Mr. 
Airy, to ask his opinion, as it was of importance that the 
clock should be the very best that the science and skill of the 


eouniry can su^ly, whether the best means of obtaining 
such a clock would be to call upon some of the most emi- 
nent clockmakers to send in specifications^ drawings^ and 
estimates, of the clock which each would recommend and 
would be prepared to make ; or whether it would be better 
to place the matter in the hands of some one experienced 
mechanician, and to adopt the description of clod which 
he might recommend, leaving the execution of it open to 
tender ; and he requested the astronomer royal to send him 
the names of the persons or person he would recommend 
for the service in either case. 

Mr. Airy repKed that a nearly sinular question was ad- 
dressed to him in 1843 by the Graham Committee with 
respect to the Exchange clock, and that he then replied 
that certain conditions ought to be laid down, which he 
would furnish, and that the clockmakers^ plans should also 
be submitted to him for his opinion, and that the Commit- 
tee should refer to him for a certificate at the completion of 
the work: that those conditions were adopted by that 
Committee; and that the result is that a clock has been 
put up, which is superior even to most astronomical clocks, 
and possesses these rare advantages, that the first stroke of 
each hour is correct as to time within less than one second, 
and that a person standing on the pavement can take time 
from the face without an error of a second. Mr. Airy 
therefore proposed that a similar course should be followed 
here. He says that he suggested to the Gresham Commit- 
tee the names of Mr. Vulliamy and Mr. Whitehurst; but, 
by arrangements with which he was not acquainted, the work 
was placed in the hands of Mr. Dent, chronometer-maker 


(f . e, before that time not a tuiret-clock maker) ; and he 
was bonnd to- say that Mr. Dent had carried out his views 
most completely^ making in the mechanical arrangements 
which he (Mr. Airy) had suggested some judicious altera- 
tions^ which received his entire approval. Under all the 
circumstances^ considering that a new clock pretending to a 
degree of accuracy equal or superior to that of the Ex- 
change^ must probably contain some of Mr. Dent^s inven- 
tions^ and would at any rate be improved by his experience, 
that the trust is^ so to speak^ confidential, and that there 
is no such thing as a market for such clocks — ^Mr. Airy 
thought it would probably be the best course to transmit 
proposals (including his enclosed conditions) to Mr. Dent, 
and to ask for his tender; and if his price should not 
be excessive, that he should be employed : if it appeared 
objectionable, other makers should be apphed to ; but he 
thought only the two he had named. 

The Board did not however adopt Mr. Air/s recom- 
mendation to apply in the first instance to Mr. Dent ; but 
Mr. Barry a few days afterwards, iti July, 1846,* wrote to 
Mi*. Vulliamy to inform him that the Board had deter- 
mined to invite the tender of plans and specifications from 
.different quarters for the clock ; to be based upon the en- 
closed general conditions prescribed by the astronomer 
royal, who (Mr. Barfy stated) had recommended that ap- 
pHcation should be made to Mr.; Dent and Mr. Whitehurst 
as well' as to Mr. VuUiamy. He also informed Mr. V. 
that if he should have acquired any information which 

* It may be observed that just at this time Lord Canning re- 
signed, and was succeeded by the present Lord Carlisle. 

MR. airt's conditions. 257 

might lead him to suggest a departure from the enclosed 
conditions he was at liberty to do so in sending in his 
plans. And further that the Board considered it conve- 
nient that he should submit the estimated cost of supplying 
and completing the clock in all respects according to the 

186. The following are Mr. Air/s 'Conditions to be 
observed in regard to the construction of the clock. 
; '1. The clock frame is to be cast-iron, and of ample 
strength. Its parts are to be firmly bolted together ; where 
.there are broad bearing surfaces, these surfaces are to be 

' 2. The wheels are to be of hard bell- [gun-] metal, with 
.steel* spindles working in bell-metal bearings, and proper 
holes for oiling the bearings.* The teeth of the wheels are 
to be cut to form on the epicycloidal principle : [nothing is 
jsaid about the pinions, of what shape or material they are 
.to be.] 

/3. The wheels are to be so arranged that any one can 
be taken out without disturbing the others. 

/ 4. The pallets are to be jewelled. 

'5. The escapement is to be dead-beat, or something 
equally accurate, the recoil escapement being expressly 
excluded. ' 

' 6. The pendulum is to be compensated. 

* This is a mistake : such holes are unnecessary and mischievous, 
as they will let in dust, which will do more harm to the pivots than 
want of oil would, and clock pivots can be oiled perfectly well without 
lioles. In heavy machinery going with considerable speed, and re- 
quiring a great deal of oil, the case is different. 


' 7. The train is to have a remontoire action^ so con- 
structed as not to interfere with the dead beat principle of 
the escapement. 

' 8. The clock is to have a going fusee [barrel. No 
particular kind of going barrel is here specified, but it 
appears from the subsequent papers that one similar to 
that in the Exchange clock is intended.] 

^ 9. It will be considered an advantage if the external 
minute-hand has a discenuble motion at certain definite 
intervals of time. 

^ 10. A spring apparatus is to be attached, for accele- 
rating the pendulum at pleasure during a few vibrations: 
[this wiU be explained presently.] 

'11. The striking machinery is to be so arranged that 
the first blow for each hour shall be accurate to a second 
of time. 

' 12, and 13, [relate only to a possible electrical con- 
nexion with the Greenwich Observatory for the purpose of 
making the clock report its own behaviour to the astrono- 
mer royal.] 

' 14. The plans before commencing the work, and the 
work when completed, are to be subjected to the approval 
of the astronomer royal. 

'15. In regard to articles 5 to 11, the maker is re- 
commended to study the Exchange clock.' 

188. An explanation will probably be required of the 
apparatus referred to in the tenth of these conditions. It 
is evident that a clock with a two seconds pendulum cannot 
be altered by any less amount than two seconds, without 
handling the pendulum in a manner which is both difficult 


and unsafe with a heavy pendulum. Mr. Airy therefore 
contrived for the Exchange clock an apparatus which 
enables it to be set to any fraction of a second. It con* 
sists of a long spring set upright upon a frame which slides 
under the pendulum bob, and is so arranged that by 
pulling a string in the clock room the frame can be brought 
into such a position that the spring hits the pendulum at 
every swing in one direction, which accelerates its vibration 
a little. Therefore, if the clock is a few seconds too slow, 
the man merely puUs up the string, and holds it till he 
observes the second-hand and the beat of the clock agree 
exactly with that of the chronometer in his hand. If it is 
a few seconds (not an even number) too fast, it is first put 
back one beat too much, and then accelerated to make it right. 
I may observe here, that a clock should never be altered 
by taking hold of the pendulum anywhere but at the bob ; 
and indeed it is better not to meddle with the pendulum at 
all, but to put the clock back by holding the scape-wheel 
carefully for as many beats as may be necessary, if it is too 
fast ; and if it is too slow, by first putting the hour wheels 
forward by their adjusting work (150) and then stopping 
the scape- wheel; and you may put the scape- wheel back the 
time of two beats in one, if instead of merely holding it 
steady you make it escape the wrong way. In regulators, 
which always have a second-hand, the best way of retard- 
ing the clock a few seconds is to put your finger firmly on 
the seconds dial, just before the hand when it is deady so as 
to stop it for the proper number of seconds. If it is less 
than a minute too slow, you must first put the minute-hand 
forward a minute, and then stop the second-hand. 


188. At this point the Westminster clock correspond- 
ence begins to assume rather more of personal than of horo- 
logical interest, of which however I shall divest it as far as 
possible. But it is necessary to the understanding of the 
matter to state that Mr. Vulliamy answered Mr. Barr/s 
letter, enclosing these conditions, by declining to enter 
into a competition, chiefly on the ground that he objected 
to the astronomer royal as sole referee, because he con- 
sidered other individuals as well, if not better qualified to 
offer an opinion on the subject ; and secondly, because Mr. 
Airy had shown himself prejudiced in favour of Mr. Dent, 
by having publicly stated, through the Gresham Committee, 
that he ^had no doubt the Exchange clock was the best 
public clock in the world/ and in a subsequent letter of 
March, 1847, to the same effect, he refers also to a letter 
of Mr. Air/s to Mr. Dent, saying, ^ I shall state without 
hesitation that I consider you the most proper person to 
be entrusted with the construction of another clock of simi- 
lar pretensions.^ On which it is obvious to remark, that 
whether the recommendation, of a man to make a second 
piece of machinery, because he had already made one of the 
same kind to the satisfaction of his employers or their 
referee (who had riot recommended him) is what is com- 
monly understood by the word ^ prejudice,^ or not, Mr. 
Dent had got no benefit from Mr. Air/s recommendation 
that he should be employed, because the Board did not 
adopt it; and he thenceforth stood in exactly the same 
position as Mr. Vulliamy and Mr. Whitehurst. 

In the- following month Mr.. VuUiamy sent to the 
Board, through Mr. Barry, his drawings and specifications; 


adding, however, that he had not prepared any estimate, 
for the reasons stated in his previous letter on that subject, 
and for the further reason that it would be useless, since 
he had declined to make the clock under the direction of 
the astronomer royal; and concluded by thanking Mr. 
Barry for the very honourable and friendly manner in which 
he had been treated by him throughout the business. 

Mr. Vulliamy also, at the same time, availed himself 
of Mr. Barr/s invitation to offer such suggestions as oc- 
curred to him for departing from any of Mr. Air/s con- 
ditions, and suggested a departure from no less than half 
of them ; and, in fact, did not assent to one of them which 
was not in accordance with his own previous practice. Mr. 
Vulliam/s description of his own plans occupies 27 folio 
pages; and therefore, though they might be interesting to 
clockmakers, it is impossible to make any use of them here, 
or, I may add, of the othqr specifications, descriptions, &c., 
on account of their length. Mr. Air/s opinion of them 
all I shall have to state presently. 

There was some farther correspondence about a card- 
board model, which Mr. Dent offered to furnish by way of 
illustrating hi^ plans, but which Mr. Vulliamy rightly said 
was only throwing away time and money. In the letter in 
which he declined to furnish such a model, he repeated his 
objections to Mr. Airy, and added, that in several public 
cases of reference of horological inventions the reference 
was not to an individual but to a committee. Instead of 
the card-board model, Mr. Vulliamy informed the Board 
that he was then making for Mr. Feto a quarter clock of 
rather more than one fourth of the size of the great clock. 


and that he was purposely making it as like the great clock 
as was practicable. I have myself seen this clock since it 
was finished^ and it is a very handsome and well-executed 
piece of machinery ; and in like manner the Exchange clock 
might be looked upon as Mr. Dent^s models though not 
intended to be exactly followed. 

Various communications took place between Mr. Airy, 
and Mr. Whitehurst and Mr. Dent, respecting the details of 
the clock, and finally they both sent in their tenders, draw- 
ings, fee, which were, together with Mr. Vulliam/s plans 
and observations, submitted to Mr. Airy; and he also went 
to make a personal inspection of Mr. Dent and Mr. White- 
hursf s factories, and in May, 1847, reported to the Board 
to the following effect. That as regards their factories and 
tools, either of them, with some assistance from an engineer's 
establishment for the large frame and the great wheels, is 
competent to undertake the work. That Mr. Dent's expe- 
rience, previously to his commencing the manufacture of 
turret clocks when he undertook that of the Exchange, had 
been chiefly in astronomical clocks and chronometers, in 
which he had been compelled to pay the utmost attention 
to the excellency of fine workmanship and to secure great 
accuracy of results; and that since he commenced the 
manufacture of turret clocks he appears to have entered in 
an enterprising manner into that business, examining the 
construction of foreign clocks of celebrity, and making 
himself acquainted with the hterature of the subject. That 
Mr. Whitehurst has had very considerable experience in 
the manufacture of turret clocks, and is enthusiastically 
fond of clockmaking ; but he has seen none but English 

MR. AIRr's REPORT. 268 


clocks, and those principally in a limited district. That if 
it were necessary to entrust the making of the clock with- 
out any control to one or other of them he should prefer 
Mr. Dent; because he thinks it easier for him to acquire 
Mr. Whitehurst^s solidity than for Mr. Whitehurst to 
acquire Mr. Dent^s accuracy; but that under the most 
trifling control, either of them will certainly construct the 
clock in a perfectly satisfactory way. Lastly, he notices 
the two estimates, which were, Whitehurst, £3373 ; Dent, 
£1600. He says that it is out of his power to explain the 
astonishing difference between them. It is not of much 
consequence to the public what the explanation is; but 
Mr. Airy suggests, first, that Mr. Dent may really be able 
to do the work at less cost to himself than Mr. Whitehurst 
(I suppose from his previous experience in the Exchange 
clock) ; and secondly, that he may be willing to construct 
the clock, even at a loss to himself, for the sake of the 
reputation which he hopes to acquire by the makiug of 
such a clock, while Mr. Whitehurst has made his estimate 
at what is called a paying price. Mr. Airy (naturally 
enough after what had occurred before) declined to offer 
any further suggestion as to which of the two candidates 
should be employed. 

He added in a separate letter some remarks upon Mr. 
Vulliamy^s plans and papers, as they had been submitted to 
him by the Board with the others, although he considered 
it impossible that Mr. V. could be employed to make the 
clock, as he refused to comply with the proposed conditions. 
He says that in regard to the provisions for strength, soli- 
dity, power, and general largeness of dimensions, the plans 



are excellent ; but that in delicacy they fail^ and fail so 
much that he considers that such a clock (except of course 
as to its size) would be a village clock of very superior cha- 
racter^ but would not have the accuracy of an astronomical 
clock. The meaning of which is, that there are no provi- 
sions in the clock proposed by Vulliamy for securing greater 
accuracy of going, striking, or indicating the time, than in 
a village clock of the ordinary construction but of superior 
workmanship. With regard to the personal objections to 
himself, Mr. Airy says that Mr. VuUiam/s demand for a 
committee is not borne out by the instances he had cited; 
because in all those cases the question was about the intro- 
duction of principles, which if established were to be applied 
to an infinite number of instances ; whereas here there is no 
new principle : the instance is unique : its effects are only 
those of the display of a good specimen of the present state 
of the art : that (besides the similar case of the Exchange 
clock) Mr. Airy is often requested by persons in want of 
chronometers to recommend makers of them ; and, finally, 
that Mr. Whitehurst, a bond fide competitor, had made no 
objection to him as referree : in fact, he acknowledges 
in his letters to the Board, that he had received valuable 
information from the astronomer royal. 

Since this report upon the plans in May, 1847, nothing 
more has been done about the matter, except that on the 
Board being reminded of their letter to Mr. Dent, Mr. 
Barry was stopped from ordering any more of the smaller 
clocks of his own authority, as he had done for the House 
of Lords. 

Many other remarks on the personal questions that have 


been very unnecessarily introduced will have occurred to 
the readers of the correspondence, or even of this abstract 
of it ; but as they are of no importance to the science 
of horology, I shall leave the reader to make them for him- 
self. I will only add, for the information of those who 
have no other means of knowing it, that none of the 
gentlemen proposed by Mr. Vulliamy as referees have 
given, to the public at least, any reason for believing that 
they have paid any particular attention to the subject of 
clockmaking, which, as to all the most important parts of a 
dock, is a perfectly different thing from engineering. And 
though I do not agree with the astronomer royal as to some 
of the details of the plans suggested or approved by him, I 
have no hesitation in saying that, unless his ^general condi- 
tions^ are substantially followed, the clock will not be what 
it ought to be, and what both the First Lords of the "Woods 
and Forests have declared that it was intended to be. In- 
deed, imless it is so made, it will be a mere waste of money 
to make it at all, as there are plenty of clocks in that neigh- 
bourhood for aU the ordinary purposes of public clocks, 
though they axe of no use for the extraordinary purposes 
for which this clock is intended, and is really wanted, and 
which it will answer if properly made, but not otherwise. 

It is satisfactory to be able to add that, if it is really to be 
ever made, and made properly, there wiU be no reason 
to regret the delay that has taken place, because in the 
mean time some experience has been gained which wiU 
enable a better clock to be constructed than that con- 
templated in the plans which were settled three years ago^ 
and at quite as little expense. 




189. The gratification of the curiosity of the readers of 
this book on the subject of the Westminster clock was not 
the only object of giving this history of it. I wish also to 
point out to individuals and pubUc bodies who want to pro- 
cure really good turret clocks a few things which they 
ought to attend to^ but hardly ever do. Of course^ very 
few such clocks as the Westminster or the Exchange clock 
are wanted ; but as much as three or four hundred pounds 
is sometimes spent upon a clock for a large town^ which is, 
after aU, not eqnal to the cheapest ^regulator/ whicli may 
be bought for 85 guineas. Instead of spending a large 
sum of money upm such a clock, if not the best than can 
be made, it would be in every way better to get a large 
and strong clock for half the money, and a preUy good 
regulator with it, and make the clockmaker who has the 
care of it set it by the regulator every day if need be; or 
what would be still better and cheaper, a dipleidoscope 
(9), by which the clock can be corrected independetUlf 
every day when the sun shines at noon. The truth is, that 
what the astronomer royal said of the great dock may be 
said with nearly the same truth of clocks very far short of 
that in accuracy— *^ there is no market for such clocks.' 
Plenty of clockmakers will contract for them and make 
them, no doubt fairly and honestly enough as regards 
the quantity of labour expended on them; but very few in* 
deed,-*^o little demand is there for such things, — ^really 
know all the particulars on which labour is worth expending, 
and on which it is not. 


There is a very sensible remark in one of Mr.VulKam/s 
letters (whicli it was unnecessary to refer to for any other pur- 
pose) to tlus effect that a great deal of superflnous work is often 
expended in polishing parts of the clock which have nothing 
to do, as indeed I have already noticed in § 181. But it 
requires some boldness in a clockmaker to omit these 
things, for nine people out of ten who go to look at 
a clock judge of its goodness merdy by its finish (which in 
small work, especially watches, is generally not a bad test, 
by a sort of convention among the makers) ; and even the 
tenth person, though he may be aware that the going of a 
clock does not depend on the lacquering of the brass or the 
polishing of the ironwork, does not know what it really does 
depend upon. How many people, for instance, know 
whether a scape-wheel ought to be light or heavy — 
whether the teeth ought to fall very near the comer of the 
pallets, or a good way up on the dead part— whether 
a pendulum is better fixed close behind the pallet-arbor, or 
on some convenient part of the wall at the side of the 
dock — ^why some clocks will go very well with short pen- 
dulums, and others go much worse with long and heavy 
ones— whether cast iron cams on the great striking wheel 
cause more or less friction, and more or less waste of power 
than polished steel pins on the great wheel or on the second 
wheel— and why in one case cast iron wheels will act with 
less ftiction than cut brass wheels in the other case-and 
so forth ? 

There can be little doubt therefore that for obtaining 
a really good public clock, such as most large towns have 
paid for, whether they possess it or not, the only safe way 

N 2 


is to go to some one^ or at the most some tliree or four 
makers of the first reputation^ and adopt the clock which 
is proposed by the one whom you ultimately select, either 
with reference to their price or other considerations. Of 
course the more ordinary the clock is required to be, the 
larger will be the number of persons competent to make it. 
And where tenders are obtained, from any but snch a select 
number of makers, as I have just now supposed, it is espe- 
cially necessary that the advice of some competent person 
should be obtained either as to the conditions to be ob- 
served, or as to the character of the clocks usually made 
by the persons proposed to be employed, as there is always 
a strong family resemblance between the clocks of the same 

190. With regard however to testimonials respecting 
clocks I must caution churchwardens and others to whom 
they are sent, that they are (like all other testinK)nialB 
nowadays) in many instances most fallacious, being fre- 
quently given by persons who have no means of ascertaining 
the real rate or the real value of the clocks about which 
they testify, and who are quite incompetent to fonn aiL ac- j 
curate judgment of the construction of a clock. Two 
opposite instances occur to me, as illustrations of the Talne 
of testimonials of this sort. I have seen a printed tes- 
timonial about a clock, to the effect that it had not varied 
five minutes in several months. Now such a tesdmoiiial 
is not worth five farthings ; for it may have meant, eith« 
that the clock had a steady gaining or losing rate of twc 
or three seconds a day, which amounted in several monthf 
to five minutes; or that it was never more than two nunutei 


too slow or three minutes too fast; or (what was probably 
the &ct) something between the two : in the first case the 
dock was a very good one; in the second case it was a 
very bad one ; and in the third it may have been anything 
between those two extremes. On the other hand, I heard 
of a clergyman in London being asked the character of a 
clock that had been made for his church by an eminent 
maker, and he replied that it went very ill : this came to 
the ears of the person most interested in the matter, who 
from his own or his men's periodical observation of it 
thought it deserved a much better character, and he re- 
quested the clergyman to inform him by what standard of 
time he was in the habit of trying it ; and he replied, ' by 
the clock of the principal church of the parish.' And it 
being thought by those who had made the original 
inquiry that the clocks of chapels of ease are not bound 
by ecclesiastical obedience to the mother church, and other 
inquiries proving satisi^tory, the maker of the contu* 
maeious clock was employed Upon the business in question. 
I can give one hint on this subject which may be useful, 
viz : that any testimonial about the going of a clock for 
so many months ought to extend from winter to summer, 
or it is of no real value. 

191. With regard to clocks of what be called the 
second degree of excellence, or, as we may for convenience 
call them, village clocks, there is of course not quite the 
same necessity for caution as to the admission of can- 
didates. At the same time nothing can be worse than 
the common practice of churchwardens, government offices, 
and public bodies, of issuing a notice that a clock is re- 


quired to show time on a face of such a size, and to strike 
on a bell of such a weight, to go eight days, all the wheels 
to be of the best brass, and several other best things, and 
that tenders must be sent in by 12 o'clock on sach a day^ 
the only condition they seem to consider of iniportanoe. 
I said nothing can be worse ; but I belieye there is one 
other worse method, which I lately heard of hdng adopted 
by a certain town council, who after spending several 
thousand pounds on new markets, naturally thought that a 
small clock on the building would be useful to those who 
frequent it : go to some respectable maker and ask what he 
will make the clock for ; then take his estimate about to 
other persons and inquire for how much less ihej will do 
it. In this way you may make sore of getting a dock 25 
per cent, cheaper and 50 p» cent, worse than the original 

As an illustration of the results of the system of ten- 
dering without a detailed specification, I may be allowed to 
relate an anecdote in which I was accidentally concerned 
not long ago. I was staying at a place where the in- 
habitants were about to put up a church dock; and it 
being known to the clergyman that I took some interest in 
such things, I was requested to attend, as a sort of as- 
sessor, at a parish meeting at which the selection of the 
maker was to be determined. They had received six or 
seven tenders, varying in amount from £60 to 100 guineas ; 
and it appeared that several of those who had sent in 
tenders also proposed to tendesr themselves to give any in- 
formation that might be required by the meeting. The 
parishioners said they did not know what information to 


ask for ; all the candidates offered to make the best pos- 
sible clocks and several of them had sent testimonials 
equally good as to their capability. It appeared therefore 
that the only way in which I could help them was to en- 
deavour to ascertain from the clockmakers who were in 
att^dance what kind of clock they really intended^ and 
were able to make. The result was (as I afterwards 
heard)* that two of the candidates^ who were prepared for 
an examination by the vestry^ declined the examination by 
their assessor ; and also that* I had no difficulty in deciding 
that the intended clock of one of the two who did appear 
would be dear at any price^ and in selecting another as 
competent to do the work^ and intending to do it in such 
a manner as would be creditable to himself and satisfactory 
to those who had to pay for it; and he agreed to make it 
according to certain conditions which I was to famish him 
with. I have lately had an opportunity of seeing the 
clocks and I was glad to find my selection justified by the 
result. Of course I do not mean to say that some of the 
other makers who sent in tenders^ but did not come to the 
place^ might not have done it just as well; but I relate 
this anecdote to show that out of six or seven persons who 
all professed to make a clock with everything of the best 
quality^ there were at least three who were either not able 
or did not intend to do what they promised^ and yet were 
just as Ukely in the common course of things to have been 
employed, as the person who was employed ; in fact more 
likely, because his tender was very nearly the highest. 

The truth is that the persons who prepare what they 
call the ^ecificatian for a public clock generally do not 


know any more than any man in the street what ther 
really want^ or ought to want : they know the remit they 
want^ if they have a bell or a clock face ready made^ but 
nothing more. And it is quite a mistake to impute dis- 
honesty to any clockmaker^ merely because he sends in 
either a very high tender or a very low one ; for until they 
are examined by some competent person^ there is no means 
of knowing whether the clock which each of them proposes 
is not really a fair clock to make for the money he ssks ; 
and one is just as much in conformity with the specifica- 
tion, which specifies nothings as the other. And the prac- 
tical result is, that the best makers will not take tde 
trouble to tender, as they are sure to be underbid. Every 
now and then the architect tries his hand at the clock spe- 
cification. But even architects are not omniscient. I have 
seen a specification — ^and a second explanatory specification 
too — ^furnished sometime ago by an eminent architect for m 
important public clock, which, if it had been printed instead 
of shown to me in manuscript, I would have copied here, 
not the least by way of any reflection on the gentleman who 
wrote it, but by way of showing how necessary it is that 
the public should resort to some better way of securing a 
good clock, than putting it into the hands of the architect. 
192. Still the question remains, how are people in 
ordinary circumstances, who want as good a clock as pos- 
sible for the money they can afford, to proceed to obtain it. 
If the clock is really intended to be a first-rate public time- 
keeper, such as all large towns ought to have, I have 
already mentioned what appears to be the only safe way of 
obtaining it. And I may add, that I believe there are few 


towns in the kingdom of 10,000 people, in which there 
are not to be seen a number of clocks on different public 
buildings, which, even when they do not all set up for 
themselves, only keep among them a sort of conventional 
time, quite distinct from any of the known descriptions of 
time, sidereal, solar, or mean, Greenwich, . or local time. 
And the money spent* on aU these bad clocks would have 
suiBced to procure one as good as can be made, and strik- 
ing so as to be heard all over the town, and from which 
any number of comparatively cheap dials (or silent clocks) 
might be daily regulated, if necessary. Indeed it should 
be remembered more frequently than it is, that the strikmg 
of a public clock is what people really go by, and set their 
own clocks by. There are many places in which it would 
be better, both on account of the architecture of the church 
where the beUs are, and its position, to put a striking 
clock without any external face (which moreover gives it 
considerable advantages in going) in the church, and to 
put a large dial in some other more conspicuous part of the 
town. Peterborough and Lichfield cathedrals, and several 
handsome churches which I could mention, are not defaced 
with visible dials, there being no sufficiently large space of 
blank wall on the bell towers on which dials could be 
placed. And the money that dials would cost, including 
the extra work they frequently require in the clock, may be 
much more profitably spent upon quarters, which ought to 
be more frequent than they aire. 

193. It appears to me that where the churchwardens 
have no better means of obtaining assistance, I may pos- 
sibly make this book worth the two shillings it will cost 


them, if I conclude it with some hints as to the conditioiis 
or tenns which they ought to require those who send in 
tenders for a dock to observe or to specify; making them 
in fsud, or such of them as may be finally adopted, a portion 
oi the contract with the clockmaker, and also affording the 
means of comparing one tender with another. And tiie 
maker may be required to submit to the judgm^it of some 
person conversant with machinery, whether the conditions 
have been properly observed, before he is paid. 

On the subject of the pendulmn, and indeed on all 
other matters here mentioned, I must refer to my re- 
marks in the earlier part of this chapter. As it is a mere 
question of money whether the pendulum should be 
compensated or not, the purchasers must deitermine it 
for themselves. If the pendulum is as much as 8 feel 
long (which nearly every place will admit of), the bob 
should not be less than 1^ cwt; and a 14 feet wooden 
pendulum may as easily have a bob of 2 or S cwt. as a 
lighter one, as there are no compensation tubes. A zinc 
compensation tube should be required to be made as de- 
scribed in § 132, at least until some method is found of 
rolling ziQC tubes thick and solid enough for this purpose. 

If the pendulum is not compensated it must be of 
wood, either deal or mahogany, straight in the grain^ and 
well varnished, and of the thickness before mentioned (see 
§ 138 and the following ones). 

The escapement to be what is technically called half- 
dead (38), and the pin-wheel escapement (46) to be pre- 

The frame to be of cast iron, the maker to state of what 


thickness^ and also the general construction and size of it : 
it being an express condition that the main body of it 
should never require taking to pieces for cleaning the clock 
when once fixed (136). 

State how many days the clock is to go^ and whether 
each train is to have three or four wheels, with reference 
to the circumstances mentioned in § 189, &c. In all cases 
a three wheeled train to be preferred, and the striking is 
to be from the great wheel, especially in large clocks, 
unless for some special reason it is impossible ; and de- 
scribe the cams or pins and lever arrangements. 

The pallets, and all the wheels, except the great ones, 
in each train are to take out separately; or it will be suf- 
ficient with respect to the wheels next above the great 
wheels if their bushes take out so that the wheels and 
pivots can be cleaned. 

The barrels to be of strong sheet iron brazed together, 
and, if possible, of such a size that all the available fall can 
be used for the weights hung by a double line ojily : if 
however this would make the barrek smaller than four 
inches in diameter for a small clock, and five for a lao^ 
one, there may be three lines."**" • 

The number of leaves in all the pinions to be stated, 
and of course high numbers to be preferred, except that if 
lantern pinions with steel pins are used, the numbers may 
be one-third kss than those of leaved pinions. The great 
wheel of the going part should not have to drive fewer 

* It must be remembered that the diameter of the barrels must 
be less than that which appears by oalcnlation to produce a certain 
fall, by an amount equal to the thickness of the rope. 


tban 16 leaves^ and the higher wheek in the train 12. If 
the striking is done by the second wheel, its pinion should 
not have less than 16 or 18 leaves; if not, 10 or even 9 
leaves will do. 

The size and thickness of the great wheels to be stated, 
and the size and number of teeth or pins of the scape-wheel. 
If the great wheels turn in three or four hours a foot 
in diameter is enough in ahnost any case for the going 
great wheel, and 1^ for the striking great wheel, each being 
about an inch tliick. If they turn slower, they must 
be* larger, in proportion, for large clocks. 

State whether you intend to use brass or gun-metal : if 
brass, aU the wheels to be hardened by hammering. 

State what kind of maintaining power you intend to 
apply, the preference being given to Harrisons's going 
ratchett, unless the bolt and shutter is made as described in 
§ 159. 

State the size of the bevelled wheels in the leading-off 
work, if any; no rule can be laid down for them, but they 
ought to be from about five to nine inches in diameter, 
according to the size and number of the dials; those be- 
longing to each dial (if mote than one) need not be so thick 
as those that work the whole : see § 149. These wheels 
may be of cast iron. 

All the wheels and pinions of the dial work to be of 
brass, to prevent rusting. In all other parts of the clock 
the pinions to be of steel (not iron case-hardened) (152), and 
the pivots also. 

Describe the dial you propose, if it is not already pro- 
vided; the hands to be as described in § 155. 


If there are quarters on two bells^ the striking wheel to 
be like that of the hour, with pins or cams alternately on 
each side; and in both parts the levers are not to have their 
axis between the striking pin and the wire (167). If the 
quarters are on four beUs, see § 168. State whether the hour 
is to be Ifet off by the quarters or the going part : the latter 
in very accurate clocks is the best, but requires rather more 
force in the going part. 

The hammer-shanks and tails, and cranks, to be l8 
inches long at least, where the position of the bells allow it; 
and as few cranks as possible to be used, and all the pivots 
to have brass bushes. 

State the weight and rise of the hammer-head from the 
bell, and the weight and fall of the striking weight with 
which you intend to obtgon that rise (165). 

There must be an internal regulating dial for minutes, 
adjustable by some such method as described in § 150, 
according to the size of the clock ; and either a small hand, 
or a visible mark on the scape-wheel with a fixed index, to 
examine the going of the clock to seconds, if it is intended 
to be a very good one. The regulating dial to be visible (if 
desired) through a glass in the clock-case. 

The whole to be enclosed in a wooden case, which can 
be opened on all sides, but locks up securely, leaving the 
winding holes open,* so that the clock can be wound with- 
out opening the case. The pendulum also to be enclosed, 

* In regulators (house clocks) the winding holes are frequently 
made in the glass, with brass caps, so that the case never requires to 
be opened, and moreover servants can then safelj be left to wind up 
the clock without the possibility of their meddling with thq hmls. 

278 ADVICE TO churchwardens; 

and where it is in a light place, a glass may be placed near 
the bottom to see the degree plate through. The weights, 
if required, to be boxed off for safety, and provisions to pre> 
vent accidents to be made where necessary (184). 

On the other hand, the clockmaker may fairly demand, 
in order that justice may be done to his dock when it 
is made, that the chamber in which it stands shall if 
possible be made so light that a candle shall never be 
required to examine the dock in the day time; otherwise 
the clock will never be properly cleaned. 

And it is equally due to the maker of a good clock, as 
well as for the interest of those to whom the dock bdongs, 
that a competent person should be employed to examine and 
dean it periodically ; who will probably w>t be the person 
who undertakes to do it for the least money. Some of tiie 
readers of this book wiQ know of a town which had to pay 


rather dearly for putting its church dock into the hands of 
a cheap and ignorant contractor; for having to do some- 
thing to the hammer, he leffc it striking with its edge instead 
of its face, and thereby cracked a very fine bell of 30 cwt. 

I do not propose the above conditions as perfect, or as 
necessary to be imposed upon clockmakers who from their 
known character can be trusted to do without. But I 
believe that few such clockmakers would desire to make 
any material alteration in them, as they by no means restrict 
them to any particular pattern of clock, and are in a great 
measure adopted from the practice of the best makers. And 
I have no doubt that such makers would rejoice to see 
some such test applied to themsdves and their competitors, 
as it would certainly tend to exclude from our churches and 



public buildings many pieces of machinery, which not only 
are a disgrace to them, but also discourage all attempts to 
improve the art; since no one will go on spending his time 
and money in contriving and making improved machines 
which are to be rejected for old fashioned and good for 
nothing articles made at somewhat smaller cost. 

I shall be glad if this, the only English book on the 
construction of Public Clocks, does anything towards raising 
the character of a machine, which, notwithstanding its ge- 
neral and important uses, has been strangely left; behind in 
the progress of mechanical improvement ; and which, not 
only in its history but in its own duration, connects the 
present with the past more than any other instrument in 
use : a machine of which it has been happily said,"^ 'there 
is no dead thing so like a living thing as a dock :' delibe- 
rately performing its appointed work by day and by night, 
with scarcely any interruption during the lapse of many 
generations of men, reminding them aR of thL own pass- 
ing away, and of the period when 'the great clock of Time 
will have ruQ down for ever/ 

* In ' The Old Church Clock/ by the Rev. R. Parkinson, Caaon of 
Manchester, a pleasing little book, but not on dockmaking.